From nheller at silcon.com  Sat Mar  1 17:32:32 2008
From: nheller at silcon.com (Neil Heller)
Date: Sat, 1 Mar 2008 17:32:32 -0800
Subject: [sf-perl] CGI working
In-Reply-To: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>
References: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>
Message-ID: <000f01c87c05$49a5c320$dcf14960$@com>

I looked and looked unsuccessfully to directly find the answer to my
question.  I was hoping someone here knew off the top of his head.

When a Perl-CGI program is launched, does it stay in memory as a singularity
(like J2EE) or does it instantiate anew each time it's called?




From friedman at highwire.stanford.edu  Sat Mar  1 18:09:35 2008
From: friedman at highwire.stanford.edu (Michael Friedman)
Date: Sat, 1 Mar 2008 18:09:35 -0800
Subject: [sf-perl] CGI working
In-Reply-To: <000f01c87c05$49a5c320$dcf14960$@com>
References: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>
	<000f01c87c05$49a5c320$dcf14960$@com>
Message-ID: <E9FE7BED-5ADA-43F5-B900-AF1678165DA2@highwire.stanford.edu>

If you are using mod_perl (or a similar setup for non-apache  
webservers) it will stay in memory between invocations. If you are  
just using CGI.pm and launching the program from a directory path,  
then it gets called anew each time.

The CGI spec says that the webserver will call out to the program anew  
every time. That is the same no matter what language the program is  
written in: perl, python, C, bash shell, etc. It's a really expansive  
feature, since you can do anything you can do on the machine from the  
webserver, but it's a bit slow due to the repeated instantiations. ;-)

You should check the documentation for your webserver about how it  
handles CGI scripts for further details.

-- Mike


On Mar 1, 2008, at 5:32 PM, Neil Heller wrote:

> I looked and looked unsuccessfully to directly find the answer to my
> question.  I was hoping someone here knew off the top of his head.
>
> When a Perl-CGI program is launched, does it stay in memory as a  
> singularity
> (like J2EE) or does it instantiate anew each time it's called?
>
>
>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

---------------------------------------------------------------------
Michael Friedman                     HighWire Press
Phone: 650-725-1974                  Stanford University
FAX:   270-721-8034                  <friedman at highwire.stanford.edu>
---------------------------------------------------------------------



From nheller at silcon.com  Sat Mar  1 21:17:28 2008
From: nheller at silcon.com (Neil Heller)
Date: Sat, 1 Mar 2008 21:17:28 -0800
Subject: [sf-perl] CGI working
In-Reply-To: <E9FE7BED-5ADA-43F5-B900-AF1678165DA2@highwire.stanford.edu>
References: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>	<000f01c87c05$49a5c320$dcf14960$@com>
	<E9FE7BED-5ADA-43F5-B900-AF1678165DA2@highwire.stanford.edu>
Message-ID: <000001c87c24$b5a38ac0$20eaa040$@com>

Are you saying that using plain Perl CGI programs in Apache means that the
application (really the script, in this case) will NOT stay in memory
between invocations?  That's a bummer for me.

Neil Heller



-----Original Message-----
From: sanfrancisco-pm-bounces+nheller=silcon.com at pm.org
[mailto:sanfrancisco-pm-bounces+nheller=silcon.com at pm.org] On Behalf Of
Michael Friedman
Sent: Saturday, March 01, 2008 6:10 PM
To: San Francisco Perl Mongers User Group
Subject: Re: [sf-perl] CGI working

If you are using mod_perl (or a similar setup for non-apache  
webservers) it will stay in memory between invocations. If you are  
just using CGI.pm and launching the program from a directory path,  
then it gets called anew each time.

The CGI spec says that the webserver will call out to the program anew  
every time. That is the same no matter what language the program is  
written in: perl, python, C, bash shell, etc. It's a really expansive  
feature, since you can do anything you can do on the machine from the  
webserver, but it's a bit slow due to the repeated instantiations. ;-)

You should check the documentation for your webserver about how it  
handles CGI scripts for further details.

-- Mike


On Mar 1, 2008, at 5:32 PM, Neil Heller wrote:

> I looked and looked unsuccessfully to directly find the answer to my
> question.  I was hoping someone here knew off the top of his head.
>
> When a Perl-CGI program is launched, does it stay in memory as a  
> singularity
> (like J2EE) or does it instantiate anew each time it's called?
>
>
>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

---------------------------------------------------------------------
Michael Friedman                     HighWire Press
Phone: 650-725-1974                  Stanford University
FAX:   270-721-8034                  <friedman at highwire.stanford.edu>
---------------------------------------------------------------------


_______________________________________________
SanFrancisco-pm mailing list
SanFrancisco-pm at pm.org
http://mail.pm.org/mailman/listinfo/sanfrancisco-pm


From woof at danlo.com  Sat Mar  1 21:39:42 2008
From: woof at danlo.com (Daniel Lo)
Date: Sat, 1 Mar 2008 21:39:42 -0800
Subject: [sf-perl] CGI working
In-Reply-To: <000001c87c24$b5a38ac0$20eaa040$@com>
References: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>
	<000f01c87c05$49a5c320$dcf14960$@com>
	<E9FE7BED-5ADA-43F5-B900-AF1678165DA2@highwire.stanford.edu>
	<000001c87c24$b5a38ac0$20eaa040$@com>
Message-ID: <1155995594.20080301213942@danlo.com>

Hello Neil,

Plain generic Perl CGI programs will not stay in memory between invocations.
This is a plus in some regards.

1. Any attempts to affect the next invocation (hacking) will fail, because the
program starts fresh.
2. No memory leaks (well, nothing you have to worry about).
3. It leaves memory! (If your application, is called once every 20 minutes, it
is not hogging memory!)
4. Easier to debug.
5. Widely supported.  (What server doesn't do this...)
6. If your program is called often, the harddisk sectors containing the program
are likely to stay in the OS cache, so its not that painful)
7. Less problems (sometimes).  You don't have issues with programs that run a
long time. (Not giving up database connections, memory hogs, caching issues,
un-initialized variables between calls, etc..)

---

There are other ways around this.  FastCGI (FastCGI.com) is a simple wrapper
that is easy to use (FCGI.pm/CGI.pm).  And there is mod_perl.

You may want to check out Catalyst: http://catalyst.perl.org/

-daniel



Saturday, March 1, 2008, 9:17:28 PM, you wrote:

> Are you saying that using plain Perl CGI programs in Apache means that the
> application (really the script, in this case) will NOT stay in memory
> between invocations?  That's a bummer for me.

> Neil Heller



> -----Original Message-----
> From: sanfrancisco-pm-bounces+nheller=silcon.com at pm.org
> [mailto:sanfrancisco-pm-bounces+nheller=silcon.com at pm.org] On Behalf Of
> Michael Friedman
> Sent: Saturday, March 01, 2008 6:10 PM
> To: San Francisco Perl Mongers User Group
> Subject: Re: [sf-perl] CGI working

> If you are using mod_perl (or a similar setup for non-apache  
> webservers) it will stay in memory between invocations. If you are  
> just using CGI.pm and launching the program from a directory path,  
> then it gets called anew each time.

> The CGI spec says that the webserver will call out to the program anew  
> every time. That is the same no matter what language the program is  
> written in: perl, python, C, bash shell, etc. It's a really expansive  
> feature, since you can do anything you can do on the machine from the  
> webserver, but it's a bit slow due to the repeated instantiations. ;-)

> You should check the documentation for your webserver about how it  
> handles CGI scripts for further details.

> -- Mike


> On Mar 1, 2008, at 5:32 PM, Neil Heller wrote:

>> I looked and looked unsuccessfully to directly find the answer to my
>> question.  I was hoping someone here knew off the top of his head.
>>
>> When a Perl-CGI program is launched, does it stay in memory as a  
>> singularity
>> (like J2EE) or does it instantiate anew each time it's called?
>>
>>
>>
>> _______________________________________________
>> SanFrancisco-pm mailing list
>> SanFrancisco-pm at pm.org
>> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

> ---------------------------------------------------------------------
> Michael Friedman                     HighWire Press
> Phone: 650-725-1974                  Stanford University
> FAX:   270-721-8034                  <friedman at highwire.stanford.edu>
> ---------------------------------------------------------------------


> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm





-- 
Best regards,
 Daniel                            mailto:woof at danlo.com



From darin_fisher at yahoo.com  Sat Mar  1 21:48:37 2008
From: darin_fisher at yahoo.com (Darin Fisher)
Date: Sat, 1 Mar 2008 21:48:37 -0800 (PST)
Subject: [sf-perl] CGI working
Message-ID: <863673.64477.qm@web35306.mail.mud.yahoo.com>

Yes, put simply.  It's not up to Perl, it's depend on the web server configuration not the application.

-Darin
 
Not a shred of evidence exists in favor of the idea that life is serious.
- Brendan Gill

----- Original Message ----
From: Neil Heller <nheller at silcon.com>
To: San Francisco Perl Mongers User Group <sanfrancisco-pm at pm.org>
Sent: Saturday, March 1, 2008 9:17:28 PM
Subject: Re: [sf-perl] CGI working

Are you saying that using plain Perl CGI programs in Apache means that the
application (really the script, in this case) will NOT stay in memory
between invocations?  That's a bummer for me.

Neil Heller



-----Original Message-----
From: sanfrancisco-pm-bounces+nheller=silcon.com at pm.org
[mailto:sanfrancisco-pm-bounces+nheller=silcon.com at pm.org] On Behalf Of
Michael Friedman
Sent: Saturday, March 01, 2008 6:10 PM
To: San Francisco Perl Mongers User Group
Subject: Re: [sf-perl] CGI working

If you are using mod_perl (or a similar setup for non-apache  
webservers) it will stay in memory between invocations. If you are  
just using CGI.pm and launching the program from a directory path,  
then it gets called anew each time.

The CGI spec says that the webserver will call out to the program anew  
every time. That is the same no matter what language the program is  
written in: perl, python, C, bash shell, etc. It's a really expansive  
feature, since you can do anything you can do on the machine from the  
webserver, but it's a bit slow due to the repeated instantiations. ;-)

You should check the documentation for your webserver about how it  
handles CGI scripts for further details.

-- Mike


On Mar 1, 2008, at 5:32 PM, Neil Heller wrote:

> I looked and looked unsuccessfully to directly find the answer to my
> question.  I was hoping someone here knew off the top of his head.
>
> When a Perl-CGI program is launched, does it stay in memory as a  
> singularity
> (like J2EE) or does it instantiate anew each time it's called?
>
>
>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

---------------------------------------------------------------------
Michael Friedman                     HighWire Press
Phone: 650-725-1974                  Stanford University
FAX:   270-721-8034                  <friedman at highwire.stanford.edu>
---------------------------------------------------------------------


_______________________________________________
SanFrancisco-pm mailing list
SanFrancisco-pm at pm.org
http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

_______________________________________________
SanFrancisco-pm mailing list
SanFrancisco-pm at pm.org
http://mail.pm.org/mailman/listinfo/sanfrancisco-pm



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.pm.org/pipermail/sanfrancisco-pm/attachments/20080301/6aed59a3/attachment.html 

From quinn at fairpath.com  Sun Mar  2 10:58:36 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sun, 2 Mar 2008 10:58:36 -0800
Subject: [sf-perl] CGI working
In-Reply-To: <000001c87c24$b5a38ac0$20eaa040$@com>
References: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>
	<000f01c87c05$49a5c320$dcf14960$@com>
	<E9FE7BED-5ADA-43F5-B900-AF1678165DA2@highwire.stanford.edu>
	<000001c87c24$b5a38ac0$20eaa040$@com>
Message-ID: <20080302185835.GA85635@fu.funkspiel.org>

On Sat, Mar 01, 2008 at 09:17:28PM -0800, Neil Heller wrote:
> Are you saying that using plain Perl CGI programs in Apache means that the
> application (really the script, in this case) will NOT stay in memory
> between invocations?  That's a bummer for me.

Check Apache::Registry on CPAN.  It lets you "[r]un unaltered CGI
scrips under mod_perl."

In general, it Just Works.  But there are some subtle problems that
can crop up, depending on how your CGI scripts are coded.  The
_mod_perl Developer's Cookbook_ has a longish section that explains
these (chapter 2, section 2).

In the meantime, your best bet is to use strict; and use warnings; in
every CGI script (if you don't already).  That will expose many of
the problems right away.

HTH,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From david at fetter.org  Sun Mar  2 11:04:47 2008
From: david at fetter.org (David Fetter)
Date: Sun, 2 Mar 2008 11:04:47 -0800
Subject: [sf-perl] CGI working
In-Reply-To: <20080302185835.GA85635@fu.funkspiel.org>
References: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>
	<000f01c87c05$49a5c320$dcf14960$@com>
	<E9FE7BED-5ADA-43F5-B900-AF1678165DA2@highwire.stanford.edu>
	<000001c87c24$b5a38ac0$20eaa040$@com>
	<20080302185835.GA85635@fu.funkspiel.org>
Message-ID: <20080302190447.GD10306@fetter.org>

On Sun, Mar 02, 2008 at 10:58:36AM -0800, Quinn Weaver wrote:
> On Sat, Mar 01, 2008 at 09:17:28PM -0800, Neil Heller wrote:

> In the meantime, your best bet is to use strict; and use warnings;
> in every CGI script (if you don't already).

s/CGI script/Perl program/g;

> That will expose many of the problems right away.

...as it will in all your other applications. :)

Cheers,
David.
-- 
David Fetter <david at fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter at gmail.com

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

From fred at redhotpenguin.com  Sun Mar  2 11:47:50 2008
From: fred at redhotpenguin.com (Fred Moyer)
Date: Sun, 2 Mar 2008 11:47:50 -0800
Subject: [sf-perl] CGI working
In-Reply-To: <20080302185835.GA85635@fu.funkspiel.org>
References: <003a01c84674$5ebfd500$4301a8c0@Untangle.local>
	<20080302185835.GA85635@fu.funkspiel.org>
Message-ID: <1204487273.339E23F0@di12.dngr.org>

On Sun, 2 Mar 2008 10:58 am, Quinn Weaver wrote:
> On Sat, Mar 01, 2008 at 09:17:28PM -0800, Neil Heller wrote:
>>  Are you saying that using plain Perl CGI programs in Apache means 
>> that the
>>  application (really the script, in this case) will NOT stay in memory
>>  between invocations?  That's a bummer for me.
>
> Check Apache::Registry on CPAN.  It lets you "[r]un unaltered CGI
> scrips under mod_perl."

You can also check out 
http://us.apachecon.com/us2007/downloads/mod_perl-2.0-from-cgi-fast.pdf

That talk has more details than you can swing an httpd process at :)

From rdm at cfcl.com  Sun Mar  2 14:24:36 2008
From: rdm at cfcl.com (Rich Morin)
Date: Sun, 2 Mar 2008 14:24:36 -0800
Subject: [sf-perl] use {strict,warnings}
Message-ID: <p06230969c3f0d5018640@[192.168.1.205]>

"use {strict,warnings}" are indeed very comforting to have as a
safety net for development.  I have been speculating about ways
to add similar (and perhaps stronger) tests to Ruby, using some
form of metaprogramming.

So, I'd like to hear folks' opinions on what kinds of tests they
find most (and least :-) useful, what tests they'd like to have,
how best to make them minimally annoying, etc.

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From david at fetter.org  Sun Mar  2 15:27:35 2008
From: david at fetter.org (David Fetter)
Date: Sun, 2 Mar 2008 15:27:35 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <p06230969c3f0d5018640@[192.168.1.205]>
References: <p06230969c3f0d5018640@[192.168.1.205]>
Message-ID: <20080302232735.GC24586@fetter.org>

On Sun, Mar 02, 2008 at 02:24:36PM -0800, Rich Morin wrote:
> "use {strict,warnings}" are indeed very comforting to have as a
> safety net for development.  I have been speculating about ways to
> add similar (and perhaps stronger) tests to Ruby, using some form of
> metaprogramming.

What a bummer for Ruby that it doesn't have this basic kind of
protection...over ten years after perl got it. ;)

> So, I'd like to hear folks' opinions on what kinds of tests they
> find most (and least :-) useful, what tests they'd like to have, how
> best to make them minimally annoying, etc.

Tests, if you'll pardon the Rumsfeld-ism, are for known unknowns, and
they frequently lead to a false sense of security.  It's the unknown
unknowns that come up and bite you on the butt.

Cheers,
David.
-- 
David Fetter <david at fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter at gmail.com

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

From quinn at fairpath.com  Sun Mar  2 15:28:10 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sun, 2 Mar 2008 15:28:10 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <p06230969c3f0d5018640@[192.168.1.205]>
References: <p06230969c3f0d5018640@[192.168.1.205]>
Message-ID: <20080302232810.GA86399@fu.funkspiel.org>

On Sun, Mar 02, 2008 at 02:24:36PM -0800, Rich Morin wrote:

> So, I'd like to hear folks' opinions on what kinds of tests they
> find most (and least :-) useful, what tests they'd like to have,
> how best to make them minimally annoying, etc.

Most useful (paraphrased):

- use of undefined variable (catches typos, among other things)
- variable already defined (for instance, two my $var = statements in one scope)
- couldn't find $package; did you remember to load $module?
- no such method $foo (happens at runtime, unfortunately, due to late binding)
- close called on an unopened filehandle

Least useful:

- Can't think of any. :)

Best way to make them useful:

- Wherever possible, make the analysis static.
- If you can't do perfectly reliable static analysis, heuristics are still good.
- Make some kind of prepackaged test target that runs your all your tests
  (or a chosen subset) and either flunks your program or gives it a clean
  bill of health.

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From quinn at fairpath.com  Sun Mar  2 15:46:33 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sun, 2 Mar 2008 15:46:33 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <20080302232735.GC24586@fetter.org>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232735.GC24586@fetter.org>
Message-ID: <20080302234633.GA86511@fu.funkspiel.org>

On Sun, Mar 02, 2008 at 03:27:35PM -0800, David Fetter wrote:

> Tests, if you'll pardon the Rumsfeld-ism, are for known unknowns, and
> they frequently lead to a false sense of security.  It's the unknown
> unknowns that come up and bite you on the butt.

True!  OTOH, for mere mortals like me, just making clear one's
assumptions about what the known behavior is supposed to be (and then
having a way to verify that they this behavior isn't regressing) is a
major boon.

Of course, when you stumble across an unknown unknown, you fix the bug
and write new a test case for it.  Then you continue marching backwards
into the future...

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From biztos at mac.com  Sun Mar  2 15:49:32 2008
From: biztos at mac.com (Kevin Frost)
Date: Sun, 2 Mar 2008 15:49:32 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <20080302232810.GA86399@fu.funkspiel.org>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232810.GA86399@fu.funkspiel.org>
Message-ID: <07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>

Most:

Undef vars, vars used only once.
Plus all the stuff I don't screw up anymore after years of strict  
warnings.

Least:

Use of uninitialized variables.

I also catch lots of typos by habitually pertidying as I code.  If my  
indentation suddenly gets thrown out of whack, command-Z and look for  
fat fingers.


  -- frosty

(via iPhone)


On Mar 2, 2008, at 3:28 PM, Quinn Weaver <quinn at fairpath.com> wrote:

> On Sun, Mar 02, 2008 at 02:24:36PM -0800, Rich Morin wrote:
>
>> So, I'd like to hear folks' opinions on what kinds of tests they
>> find most (and least :-) useful, what tests they'd like to have,
>> how best to make them minimally annoying, etc.
>
> Most useful (paraphrased):
>
> - use of undefined variable (catches typos, among other things)
> - variable already defined (for instance, two my $var = statements  
> in one scope)
> - couldn't find $package; did you remember to load $module?
> - no such method $foo (happens at runtime, unfortunately, due to  
> late binding)
> - close called on an unopened filehandle
>
> Least useful:
>
> - Can't think of any. :)
>
> Best way to make them useful:
>
> - Wherever possible, make the analysis static.
> - If you can't do perfectly reliable static analysis, heuristics are  
> still good.
> - Make some kind of prepackaged test target that runs your all your  
> tests
>  (or a chosen subset) and either flunks your program or gives it a  
> clean
>  bill of health.
>
> -- 
> Quinn Weaver, independent contractor  |  President, San Francisco  
> Perl Mongers
> http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
> 510-520-5217
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

From rdm at cfcl.com  Sun Mar  2 15:51:59 2008
From: rdm at cfcl.com (Rich Morin)
Date: Sun, 2 Mar 2008 15:51:59 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <20080302232735.GC24586@fetter.org>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232735.GC24586@fetter.org>
Message-ID: <p0623096ac3f0e9c16346@[192.168.1.205]>

DF> What a bummer for Ruby that it doesn't have this basic
DF> kind of protection...over ten years after perl got it. ;)

Fortunately, I can add features to Ruby via metaprogramming
of various sorts, make them dynamically configurable, etc.
And, in Rubinius, I can actually extend the interpreter to
(say) collect more information as the script is being parsed.
Unfortunately, all of this requires Real Work...


DF> Tests, if you'll pardon the Rumsfeld-ism, are for known
DF> unknowns, and they frequently lead to a false sense of
DF> security.  It's the unknown unknowns that come up and
DF> bite you on the butt.

Peculiarly, many in the Ruby camp use a similar argument to
say that built-in tests (as provided by strict and warnings)
are too low-level, so they don't catch semantic and design
errors.  "Just write appropriate tests."

My take is certain types of errors are known to be common,
so why not let the computer check for them?  Of course, the
tests should be optional, customizable, and dynamically
configurable.

FWIW, the Fortress type system has some interesting ideas.
For example, it can keep you from calculating furlongs per
fortnight and saving the result in a variable whose type is
fortnights per furlong.  I don't think I want to be required
to specify things that tightly for much of my work, but when
the code really has to work, this can provide useful checks.


QW> - Wherever possible, make the analysis static.
QW>
QW> - If you can't do perfectly reliable static analysis,
QW>   heuristics are still good.
QW>
QW> - Make some kind of prepackaged test target that runs
QW>   all your tests (or a chosen subset) and either flunks
QW>   your program or gives it a clean bill of health.

+1

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From quinn at fairpath.com  Sun Mar  2 16:13:24 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sun, 2 Mar 2008 16:13:24 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232810.GA86399@fu.funkspiel.org>
	<07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>
Message-ID: <20080303001324.GA86613@fu.funkspiel.org>

On Sun, Mar 02, 2008 at 03:49:32PM -0800, Kevin Frost wrote:
> Most:
> 
> Undef vars, vars used only once.
              ^^^^^^^^^^^^^^^^^^^

I second that.

Also useful:

- Deep recursion
- Read on closed filehandle (various)
- Cannot call method on undef value
- Various messages about string termination and quoting,
  runaway strings and regexes
- Unknown Unicode chars (I guess)

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From david at fetter.org  Sun Mar  2 16:29:44 2008
From: david at fetter.org (David Fetter)
Date: Sun, 2 Mar 2008 16:29:44 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <20080302234633.GA86511@fu.funkspiel.org>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232735.GC24586@fetter.org>
	<20080302234633.GA86511@fu.funkspiel.org>
Message-ID: <20080303002944.GD24586@fetter.org>

On Sun, Mar 02, 2008 at 03:46:33PM -0800, Quinn Weaver wrote:
> On Sun, Mar 02, 2008 at 03:27:35PM -0800, David Fetter wrote:
> 
> > Tests, if you'll pardon the Rumsfeld-ism, are for known unknowns,
> > and they frequently lead to a false sense of security.  It's the
> > unknown unknowns that come up and bite you on the butt.
> 
> True!  OTOH, for mere mortals like me, just making clear one's
> assumptions about what the known behavior is supposed to be (and
> then having a way to verify that they this behavior isn't
> regressing) is a major boon.

For people like me who are even more mortal, it's important, too.  In
an attempt to learn from past mistakes, I created this rule about Perl
code, which is that strict and warnings are mandatory all the way
through the code.  I hope that this rule prevents me from doing things
cleverly.[1]  It's worked pretty well so far...

> Of course, when you stumble across an unknown unknown, you fix the bug
> and write new a test case for it.

This isn't always a good idea.

> Then you continue marching backwards into the future...

It seems to me that people in our business need to cultivate
mindfulness, if you'll allow me a little casual Zen, along with good
habits.  That combination makes for a very tricky path to walk.

Cheers,
David.

[1]  Debugging is twice as hard as writing the code in the first
place.  Therefore, if you write the code as cleverly as possible,
you are, by definition, not smart enough to debug it.
                                                       Kernighan
-- 
David Fetter <david at fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter at gmail.com

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

From doom at kzsu.stanford.edu  Sun Mar  2 23:57:45 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Sun, 02 Mar 2008 23:57:45 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <p0623096ac3f0e9c16346@[192.168.1.205]>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232735.GC24586@fetter.org>
	<p0623096ac3f0e9c16346@[192.168.1.205]>
Message-ID: <200803030757.m237vjDL037371@kzsu.stanford.edu>


Rich Morin <rdm at cfcl.com> wrote:

> Peculiarly, many in the Ruby camp use a similar argument to
> say that built-in tests (as provided by strict and warnings)
> are too low-level, so they don't catch semantic and design
> errors.  "Just write appropriate tests."

Ah.  They used to be PHP programmers.


From doom at kzsu.stanford.edu  Mon Mar  3 00:44:13 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Mon, 03 Mar 2008 00:44:13 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232810.GA86399@fu.funkspiel.org>
	<07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>
Message-ID: <200803030844.m238iDWP037982@kzsu.stanford.edu>


Kevin Frost <biztos at mac.com> wrote:

> Least:
>
> Use of uninitialized variables.

Yeah, that's a problematic one, isn't it?  Perl will happily do type
conversions for you, but (if used in the recommended way, with strict
and warnings) it nags incessantly about undefs, rather than just
treating them as empty strings.

Though it could be, of course, that the constant need to dance around
this to silence suprious "unit" warnings is cleaning up my code in ways
I don't appreciate.



From biztos at mac.com  Mon Mar  3 03:43:56 2008
From: biztos at mac.com (frosty)
Date: Mon, 03 Mar 2008 03:43:56 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <200803030844.m238iDWP037982@kzsu.stanford.edu>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232810.GA86399@fu.funkspiel.org>
	<07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>
	<200803030844.m238iDWP037982@kzsu.stanford.edu>
Message-ID: <E61B6A6F-0118-1000-91B5-7910FA3F5032-Webmail-10013@mac.com>

I actually find it useful sometimes, but on average more annoying than useful.

The useful thing is that it reminds me to be clearer in things like printing and logging:

use strict;
use warnings;

my $nope = undef;
printf "You set it to '%s'\n", $nope;               # warns.
printf "You set it to '%s'\n", $nope || 'UNDEF';    # better.

But:

my $stuff = join( "fee", "fi", "fo", $nope );       # warns, annoying (to me).
die "$nope would be nice";                          # warns, annoying (to me).

On the one hand I wish undefined values could be auto-vivified to 'UNDEF' in warn and die, but on the other hand that's probably sqlplus poisoning my brain.

The problem, I find, with disabling warnings is that my list of annoying warnings may not be the same as my colleague's list.  That plus you have nobody to blame but yourself when you screw up.

The problem with *not* doing that is there is plenty of CPAN code that spews warnings all over your log files if you happen to have any.

-- f.


 
On Monday, March 03, 2008, at 12:45AM, "Joe Brenner" <doom at kzsu.stanford.edu> wrote:
>
>Kevin Frost <biztos at mac.com> wrote:
>
>> Least:
>>
>> Use of uninitialized variables.
>
>Yeah, that's a problematic one, isn't it?  Perl will happily do type
>conversions for you, but (if used in the recommended way, with strict
>and warnings) it nags incessantly about undefs, rather than just
>treating them as empty strings.
>
>Though it could be, of course, that the constant need to dance around
>this to silence suprious "unit" warnings is cleaning up my code in ways
>I don't appreciate.
>
>
>_______________________________________________
>SanFrancisco-pm mailing list
>SanFrancisco-pm at pm.org
>http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
>

From dat1965 at yahoo.com  Mon Mar  3 17:03:24 2008
From: dat1965 at yahoo.com (David Thompson)
Date: Mon, 3 Mar 2008 17:03:24 -0800 (PST)
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <E61B6A6F-0118-1000-91B5-7910FA3F5032-Webmail-10013@mac.com>
Message-ID: <706225.41072.qm@web55112.mail.re4.yahoo.com>

--- frosty <biztos at mac.com> wrote:
> I actually find it useful sometimes, but on average more annoying than useful.
> 
> The useful thing is that it reminds me to be clearer in things like printing and
> logging:
> 
> use strict;
> use warnings;

Does this help your situation any?

  use warnings;
  no warnings qw(uninitialized);

See http://perldoc.perl.org/perllexwarn.html
and perldiag

> my $nope = undef;
> printf "You set it to '%s'\n", $nope;               # warns.
> printf "You set it to '%s'\n", $nope || 'UNDEF';    # better.
> 
> But:
> 
> my $stuff = join( "fee", "fi", "fo", $nope );       # warns, annoying (to me).
> die "$nope would be nice";                          # warns, annoying (to me).
> 
> On the one hand I wish undefined values could be auto-vivified to 'UNDEF' in warn and
> die, but on the other hand that's probably sqlplus poisoning my brain.
> 
> The problem, I find, with disabling warnings is that my list of annoying warnings may
> not be the same as my colleague's list.  That plus you have nobody to blame but
> yourself when you screw up.
> 
> The problem with *not* doing that is there is plenty of CPAN code that spews warnings
> all over your log files if you happen to have any.



      ____________________________________________________________________________________
Never miss a thing.  Make Yahoo your home page. 
http://www.yahoo.com/r/hs

From biztos at mac.com  Mon Mar  3 19:04:05 2008
From: biztos at mac.com (frosty)
Date: Mon, 03 Mar 2008 19:04:05 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <706225.41072.qm@web55112.mail.re4.yahoo.com>
References: <706225.41072.qm@web55112.mail.re4.yahoo.com>
Message-ID: <486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>

Thanks, but that's exactly what I meant by my annoyances not being everyone's annoyances.

I think disabling warnings is a slippery slope even in a local context.  I suppose if everyone in an organization agreed that uninitialized warnings were the enemy it could make sense, but then why not take your case to the Perl community at large?

Of course sometimes you really have to disable warnings of some type locally or you'll go nuts, but I try to think of that as a last resort and limit it to the smallest possible context.

cheers

-- frosty


On Monday, March 03, 2008, at 05:05PM, "David Thompson" <dat1965 at yahoo.com> wrote:
>
>Does this help your situation any?
>
>  use warnings;
>  no warnings qw(uninitialized);
>
>See http://perldoc.perl.org/perllexwarn.html
>and perldiag
>
>> my $nope = undef;
>> printf "You set it to '%s'\n", $nope;               # warns.
>> printf "You set it to '%s'\n", $nope || 'UNDEF';    # better.
>> 
>> But:
>> 
>> my $stuff = join( "fee", "fi", "fo", $nope );       # warns, annoying (to me).
>> die "$nope would be nice";                          # warns, annoying (to me).
>> 
>> On the one hand I wish undefined values could be auto-vivified to 'UNDEF' in warn and
>> die, but on the other hand that's probably sqlplus poisoning my brain.
>> 
>> The problem, I find, with disabling warnings is that my list of annoying warnings may
>> not be the same as my colleague's list.  That plus you have nobody to blame but
>> yourself when you screw up.
>> 
>> The problem with *not* doing that is there is plenty of CPAN code that spews warnings
>> all over your log files if you happen to have any.
>
>
>
>      ____________________________________________________________________________________
>Never miss a thing.  Make Yahoo your home page. 
>http://www.yahoo.com/r/hs
>
>

From rdm at cfcl.com  Mon Mar  3 19:26:11 2008
From: rdm at cfcl.com (Rich Morin)
Date: Mon, 3 Mar 2008 19:26:11 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <E61B6A6F-0118-1000-91B5-7910FA3F5032-Webmail-10013@mac.com>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232810.GA86399@fu.funkspiel.org>
	<07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>
	<200803030844.m238iDWP037982@kzsu.stanford.edu>
	<E61B6A6F-0118-1000-91B5-7910FA3F5032-Webmail-10013@mac.com>
Message-ID: <p06230972c3f271692ff9@[192.168.1.205]>

At 03:43 -0800 3/3/08, frosty wrote:
> The problem with *not* doing that is there is plenty of
> CPAN code that spews warnings all over your log files if
> you happen to have any.

  ...
  use this;      # Look, Ma...
  use that;      # No warnings!
  ...
  use strict;
  use warnings;
  ...

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From doom at kzsu.stanford.edu  Mon Mar  3 20:04:42 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Mon, 03 Mar 2008 20:04:42 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>
References: <706225.41072.qm@web55112.mail.re4.yahoo.com>
	<486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>
Message-ID: <200803040404.m2444g8r055623@kzsu.stanford.edu>


frosty <biztos at mac.com> wrote:

> I think disabling warnings is a slippery slope even in a local
> context.  I suppose if everyone in an organization agreed that
> uninitialized warnings were the enemy it could make sense, but then
> why not take your case to the Perl community at large?

> Of course sometimes you really have to disable warnings of some type
> locally or you'll go nuts, but I try to think of that as a last resort
> and limit it to the smallest possible context.

I think I'm largely in agreement with you here. I do know
that you can shut off a class of warnings lexically, and I
hope it would be obvious that it's best to do this as
minimally as possible.

I think the question for me might be boiled down to this one code
example.  Is it better to code something like this:

  use warnings;
  use strict;

  my @phrases = (
              'The Rediscovery of Man',
              'The Instrumentality',
              'D\'Joan',
  );

  foreach my $phrase (@phrases) {
    my ($alpha, $ralpha, $blvd) = split ' ', $phrase;
    {
      no warnings 'uninitialized';
      print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
    }
  }

Or should that last loop be handled something like this:

  foreach my $phrase (@phrases) {
    my ($alpha, $ralpha, $blvd) = (' ', ' ', ' ');
    ($alpha, $ralpha, $blvd) = split ' ', $phrase;
    $alpha  ||='';
    $ralpha ||='';
    $blvd   ||='';
    print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
  }

In this particular case it's something of a wash in terms of
complexity or dangerousness, but you don't have to think as
hard about the second way, so I think that one wins.

I would guess that just leaving warnings on all the time wins
often enough that you're better off doing it that way by
default, and living with the occasional minor pain.


From doom at kzsu.stanford.edu  Mon Mar  3 20:13:32 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Mon, 03 Mar 2008 20:13:32 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <200803040404.m2444g8r055623@kzsu.stanford.edu>
References: <706225.41072.qm@web55112.mail.re4.yahoo.com>
	<486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>
	<200803040404.m2444g8r055623@kzsu.stanford.edu>
Message-ID: <200803040413.m244DW8a055734@kzsu.stanford.edu>

Joe Brenner <doom at kzsu.stanford.edu> wrote:

> I think the question for me might be boiled down to this one code
> example.  Is it better to code something like this:
>
>   use warnings;
>   use strict;
>
>   my @phrases = (
>               'The Rediscovery of Man',
>               'The Instrumentality',
>               'D\'Joan',
>   );
>
>   foreach my $phrase (@phrases) {
>     my ($alpha, $ralpha, $blvd) = split ' ', $phrase;
>     {
>       no warnings 'uninitialized';
>       print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
>     }
>   }
>
> Or should that last loop be handled something like this:
>
>   foreach my $phrase (@phrases) {
>     my ($alpha, $ralpha, $blvd) = (' ', ' ', ' ');
>     ($alpha, $ralpha, $blvd) = split ' ', $phrase;
>     $alpha  ||='';
>     $ralpha ||='';
>     $blvd   ||='';
>     print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
>   }
>
> In this particular case it's something of a wash in terms of
> complexity or dangerousness, but you don't have to think as
> hard about the second way, so I think that one wins.
>
> I would guess that just leaving warnings on all the time wins
> often enough that you're better off doing it that way by
> default, and living with the occasional minor pain.
>

Sorry, that second example was supposed to be like this:

  foreach my $phrase (@phrases) {
    my ($alpha, $ralpha, $blvd) = split ' ', $phrase;
    $alpha  ||='';
    $ralpha ||='';
    $blvd   ||='';
    print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
  }


(I'd been experimenting with split for a minute there, making
sure it behaved the way I thought it did...  You can not
initialize the later variables first, and expect them to stay
untouched by split if there aren't enough items returned: it
always defines the later items as undef.)


From biztos at mac.com  Tue Mar  4 09:18:31 2008
From: biztos at mac.com (Kevin Frost)
Date: Tue, 4 Mar 2008 09:18:31 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <200803040413.m244DW8a055734@kzsu.stanford.edu>
References: <706225.41072.qm@web55112.mail.re4.yahoo.com>
	<486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>
	<200803040404.m2444g8r055623@kzsu.stanford.edu>
	<200803040413.m244DW8a055734@kzsu.stanford.edu>
Message-ID: <A435CFBC-A2B1-4643-85AB-8DD5C7CD4B03@mac.com>

That's a good example.  I would say you can rewrite it so it throws no  
warnings and is still reasonably elegant, but I would use the first  
version (no warnings) before using the second (gratuitous ||=).

The warnings are telling you that your assumption there are three  
parts to each phrase is dubious.  That's a useful warning, though it's  
arguably a nit-picking sort of usefulness.

You could do something like this:

     # pretty ugly, but it works and the print didn't change:
     my ( $alpha, $ralpha, $blvd ) =
       map { defined $_ ? $_ : '' } split( /\s+/, $phrase, 3 ), '',  
'', '';
     print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";

...or like this:

     my $i = 0;
     print join( "\n", map { $i++; "$i: $_"; } split( /\s+/,  
$phrase ) ), "\n\n";

...or like this if you don't care about 4+:

     my $i = 0;
     print join( "\n", map { $i++; "$i: $_"; } split( /\s+/, $phrase,  
3 ) ),
       "\n\n";

...but if you don't like those approaches then I think the no-warnings  
bit is in exactly the right spot.

cheers

-- f.

On Mar 3, 2008, at 8:13 PM, Joe Brenner wrote:

> Joe Brenner <doom at kzsu.stanford.edu> wrote:
>
>> I think the question for me might be boiled down to this one code
>> example.  Is it better to code something like this:
>>
>>  use warnings;
>>  use strict;
>>
>>  my @phrases = (
>>              'The Rediscovery of Man',
>>              'The Instrumentality',
>>              'D\'Joan',
>>  );
>>
>>  foreach my $phrase (@phrases) {
>>    my ($alpha, $ralpha, $blvd) = split ' ', $phrase;
>>    {
>>      no warnings 'uninitialized';
>>      print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
>>    }
>>  }
>>
>> Or should that last loop be handled something like this:
>>
>>  foreach my $phrase (@phrases) {
>>    my ($alpha, $ralpha, $blvd) = (' ', ' ', ' ');
>>    ($alpha, $ralpha, $blvd) = split ' ', $phrase;
>>    $alpha  ||='';
>>    $ralpha ||='';
>>    $blvd   ||='';
>>    print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
>>  }
>>
>> In this particular case it's something of a wash in terms of
>> complexity or dangerousness, but you don't have to think as
>> hard about the second way, so I think that one wins.
>>
>> I would guess that just leaving warnings on all the time wins
>> often enough that you're better off doing it that way by
>> default, and living with the occasional minor pain.
>>
>
> Sorry, that second example was supposed to be like this:
>
>  foreach my $phrase (@phrases) {
>    my ($alpha, $ralpha, $blvd) = split ' ', $phrase;
>    $alpha  ||='';
>    $ralpha ||='';
>    $blvd   ||='';
>    print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
>  }
>
>
> (I'd been experimenting with split for a minute there, making
> sure it behaved the way I thought it did...  You can not
> initialize the later variables first, and expect them to stay
> untouched by split if there aren't enough items returned: it
> always defines the later items as undef.)
>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm


From biztos at mac.com  Tue Mar  4 09:26:04 2008
From: biztos at mac.com (Kevin Frost)
Date: Tue, 4 Mar 2008 09:26:04 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <A435CFBC-A2B1-4643-85AB-8DD5C7CD4B03@mac.com>
References: <706225.41072.qm@web55112.mail.re4.yahoo.com>
	<486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>
	<200803040404.m2444g8r055623@kzsu.stanford.edu>
	<200803040413.m244DW8a055734@kzsu.stanford.edu>
	<A435CFBC-A2B1-4643-85AB-8DD5C7CD4B03@mac.com>
Message-ID: <87527184-8708-486D-A336-94BD2E229ED6@mac.com>

Sorry, I forgot how short the e-mail elves cut their lines.  Previous  
code, shorter width:

# You could do something like this:

# pretty ugly, but it works and the print didn't change:
my ( $alpha, $ralpha, $blvd ) =
   map { defined $_ ? $_ : '' }
   split( /\s+/, $phrase, 3 ), '', '', '';
print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";

# ...or like this:

my $i = 0;
print join( "\n", map { $i++; "$i: $_"; } split( /\s+/, $phrase ) ),
   "\n\n";

# ...or like this if you don't care about 4+:

my $i = 0;
print
   join( "\n", map { $i++; "$i: $_"; } split( /\s+/, $phrase, 3 ) ),
   "\n\n";



From biztos at mac.com  Tue Mar  4 09:46:09 2008
From: biztos at mac.com (Kevin Frost)
Date: Tue, 4 Mar 2008 09:46:09 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <p06230972c3f271692ff9@[192.168.1.205]>
References: <p06230969c3f0d5018640@[192.168.1.205]>
	<20080302232810.GA86399@fu.funkspiel.org>
	<07D10AFC-5221-42A8-8C79-2A3C962EEA13@mac.com>
	<200803030844.m238iDWP037982@kzsu.stanford.edu>
	<E61B6A6F-0118-1000-91B5-7910FA3F5032-Webmail-10013@mac.com>
	<p06230972c3f271692ff9@[192.168.1.205]>
Message-ID: <F68F1BC1-9A63-49FE-8F49-15F933021D14@mac.com>

That's a cool trick but it only works if the underlying module doesn't  
use warnings.

--snip (Foo.pm)--
package Foo;

use strict;
use warnings;

sub bar {

     my ( $class, $val ) = @_;
     return "you said '$val'";
}

1;
--snip (Bar.pm)--
package Bar;

use Foo;

sub baz {

     my $class = shift;

     my $result = $class->_sloppy_operation;
     return Foo->bar($result);

}

sub _sloppy_operation {

     return;

}

1;
--snip--

Then, alas, this throws a warning:

use Bar;

use strict;
use warnings;

my $x = Bar->baz();

-- frosty

On Mar 3, 2008, at 7:26 PM, Rich Morin wrote:

> At 03:43 -0800 3/3/08, frosty wrote:
>> The problem with *not* doing that is there is plenty of
>> CPAN code that spews warnings all over your log files if
>> you happen to have any.
>
>  ...
>  use this;      # Look, Ma...
>  use that;      # No warnings!
>  ...
>  use strict;
>  use warnings;
>  ...
>
> -r
> -- 
> http://www.cfcl.com/rdm            Rich Morin
> http://www.cfcl.com/rdm/resume     rdm at cfcl.com
> http://www.cfcl.com/rdm/weblog     +1 650-873-7841

From doom at kzsu.stanford.edu  Tue Mar  4 16:05:20 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Tue, 04 Mar 2008 16:05:20 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <A435CFBC-A2B1-4643-85AB-8DD5C7CD4B03@mac.com>
References: <706225.41072.qm@web55112.mail.re4.yahoo.com>
	<486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>
	<200803040404.m2444g8r055623@kzsu.stanford.edu>
	<200803040413.m244DW8a055734@kzsu.stanford.edu>
	<A435CFBC-A2B1-4643-85AB-8DD5C7CD4B03@mac.com>
Message-ID: <200803050005.m2505KNN071981@kzsu.stanford.edu>

Kevin Frost <biztos at mac.com> wrote:

> That's a good example.  I would say you can rewrite it so it throws no
> warnings and is still reasonably elegant, but I would use the first
> version (no warnings) before using the second (gratuitous ||=).
>
> The warnings are telling you that your assumption there are three
> parts to each phrase is dubious.  That's a useful warning, though it's
> arguably a nit-picking sort of usefulness.
>
> You could do something like this:
>
>      # pretty ugly, but it works and the print didn't change:
>      my ( $alpha, $ralpha, $blvd ) =
>        map { defined $_ ? $_ : '' } split( /\s+/, $phrase, 3 ), '',
> '', '';
>      print " 1: $alpha\n 2: $ralpha\n 3: $blvd\n\n";
>
> ...or like this:
>
>      my $i = 0;
>      print join( "\n", map { $i++; "$i: $_"; } split( /\s+/,
> $phrase ) ), "\n\n";
>
> ...or like this if you don't care about 4+:
>
>      my $i = 0;
>      print join( "\n", map { $i++; "$i: $_"; } split( /\s+/, $phrase,
> 3 ) ),
>        "\n\n";
>
> ...but if you don't like those approaches then I think the no-warnings
> bit is in exactly the right spot.

It isn't so much that I don't like them, it's that working out
something like one of those solutions would've taken me a few
minutes, and worse, it probably would take someone reading the
code a few minutes to understand.

(I did have the thought if I was splitting to an array, with an
indefinite number of items, it would be time to use map.)

Like I was saying before:

> >> In this particular case it's something of a wash in terms of
> >> complexity or dangerousness, but you don't have to think as
> >> hard about the second way [leaving warnings on and
> >> manually applying a default], so I think that one wins.

Conserving mental effort is the goal, not reducing the number
of lines of code.


From biztos at mac.com  Tue Mar  4 16:25:16 2008
From: biztos at mac.com (frosty)
Date: Tue, 04 Mar 2008 16:25:16 -0800
Subject: [sf-perl] use {strict,warnings}
In-Reply-To: <200803050005.m2505KNN071981@kzsu.stanford.edu>
References: <706225.41072.qm@web55112.mail.re4.yahoo.com>
	<486A3377-0118-1000-838A-5EDAE3F22E56-Webmail-10017@mac.com>
	<200803040404.m2444g8r055623@kzsu.stanford.edu>
	<200803040413.m244DW8a055734@kzsu.stanford.edu>
	<A435CFBC-A2B1-4643-85AB-8DD5C7CD4B03@mac.com>
	<200803050005.m2505KNN071981@kzsu.stanford.edu>
Message-ID: <E7A02677-0118-1000-A239-32ADD2F40046-Webmail-10009@mac.com>


> Conserving mental effort is the goal, not reducing the number
> of lines of code.

Well put.

That's a nice thing to keep in mind when trying to measure "code kwalitee" even though it's very hand-wavy.

-- f.


From nheller at silcon.com  Wed Mar  5 15:46:51 2008
From: nheller at silcon.com (nheller at silcon.com)
Date: Wed, 5 Mar 2008 15:46:51 -0800 (PST)
Subject: [sf-perl] Can't figure this one out
Message-ID: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>

Here's an error that I can't figure out for the life of me.
The code is short; I can't figure out why the compiler thinks $action (on
the marked line) is a bare word.

The actual error message follows the source code.
I would really appreciate any help.

{
    my ($q, $ins, $action, %p, $p);

    $q = new CGI;
    $action = $q->param('action');
    $action = 'start' unless $action;

    if ($action =~ /start/i) {     <-------- line 511, the offender
        &MainPage($q, \%p);
    }
    elsif ($action =~ /new/i ||
           $action =~ /del/i ||
           $action =~ /edt/i ||
           $action =~ /save/i )
    {



[imiuser at adopt nheller]$ perl -c ad_cases.cgi
Bareword "action" not allowed while "strict subs" in use at ad_cases.cgi
line 511.
ad_cases.cgi had compilation errors.
[imiuser at adopt nheller]$ vi ad_cases.cgi
[imiuser at adopt nheller]$



From andy at petdance.com  Wed Mar  5 15:49:46 2008
From: andy at petdance.com (Andy Lester)
Date: Wed, 5 Mar 2008 17:49:46 -0600
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
Message-ID: <052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>


On Mar 5, 2008, at 5:46 PM, nheller at silcon.com wrote:

>    if ($action =~ /start/i) {     <-------- line 511, the offender
>        &MainPage($q, \%p);
>    }
>    elsif ($action =~ /new/i ||
>           $action =~ /del/i ||
>           $action =~ /edt/i ||
>           $action =~ /save/i )
>    {
>
>
>
> [imiuser at adopt nheller]$ perl -c ad_cases.cgi
> Bareword "action" not allowed while "strict subs" in use at  
> ad_cases.cgi
> line 511.
> ad_cases.cgi had compilation errors.


The "at line 511" really means "the statement that starts on line 511"  
but it looks like you have a cascading if/else chain, and so the  
ENTIRE CHAIN is "at line 511".

Break up the else/ifs and you'll get a better pointer to the offending  
line.

xoxo,
Andy

--
Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance





From sergey.fetisov at ask.com  Wed Mar  5 16:32:46 2008
From: sergey.fetisov at ask.com (Sergey Fetisov)
Date: Wed, 5 Mar 2008 16:32:46 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
Message-ID: <09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>

 

On Mar 5, 2008, at 5:46 PM, nheller at silcon.com wrote:

>>    if ($action =~ /start/i) {     <-------- line 511, the offender
>>        &MainPage($q, \%p);
>>    }
>>    elsif ($action =~ /new/i ||
>>           $action =~ /del/i ||
>>           $action =~ /edt/i ||
>>           $action =~ /save/i )
>>    {
>>
>>
>>
>> [imiuser at adopt nheller]$ perl -c ad_cases.cgi Bareword "action" not 
>> allowed while "strict subs" in use at ad_cases.cgi line 511.
>> ad_cases.cgi had compilation errors.


>The "at line 511" really means "the statement that starts on line 511"

>but it looks like you have a cascading if/else chain, and so the ENTIRE
CHAIN is "at line 511".

>Break up the else/ifs and you'll get a better pointer to the offending
line.

>xoxo,
>Andy

------------------------------------------------------------------------
------------

Yep, seems like a some line inside of the if/else statement caused the
error

By the way 

if ($action =~ /start/i) {  
        &MainPage($q, \%p);
} elsif ($action =~ /new|del|edt|save/i ) {

{

seems better to me

Sergey

From doom at kzsu.stanford.edu  Wed Mar  5 17:23:19 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Wed, 05 Mar 2008 17:23:19 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
Message-ID: <200803060123.m261NJFt094694@kzsu.stanford.edu>

nheller at silcon.com wrote:

> Here's an error that I can't figure out for the life of me.
> The code is short; I can't figure out why the compiler thinks $action (on
> the marked line) is a bare word.
>
> The actual error message follows the source code.
> I would really appreciate any help.
>
> {
>     my ($q, $ins, $action, %p, $p);
>
>     $q = new CGI;
>     $action = $q->param('action');
>     $action = 'start' unless $action;
>
>     if ($action =~ /start/i) {     <-------- line 511, the offender
>         &MainPage($q, \%p);
>     }
>     elsif ($action =~ /new/i ||
>            $action =~ /del/i ||
>            $action =~ /edt/i ||
>            $action =~ /save/i )
>     {

My experience has been that with compound booleans like this,
you're better off using more parens than you think you need:


     elsif ( ($action =~ /new/i) ||
             ($action =~ /del/i) ||
             ($action =~ /edt/i) ||
             ($action =~ /save/i) )


From doom at kzsu.stanford.edu  Wed Mar  5 17:26:16 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Wed, 05 Mar 2008 17:26:16 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
Message-ID: <200803060126.m261QGCR094802@kzsu.stanford.edu>


Sergey Fetisov <sergey.fetisov at ask.com> wrote:
> nheller at silcon.com wrote:
>
> >>    if ($action =~ /start/i) {     <-------- line 511, the offender
> >>        &MainPage($q, \%p);
> >>    }
> >>    elsif ($action =~ /new/i ||
> >>           $action =~ /del/i ||
> >>           $action =~ /edt/i ||
> >>           $action =~ /save/i )
> >>    {
> >>

> >> [imiuser at adopt nheller]$ perl -c ad_cases.cgi Bareword "action" not
> >> allowed while "strict subs" in use at ad_cases.cgi line 511.
> >> ad_cases.cgi had compilation errors.

> >The "at line 511" really means "the statement that starts on line 511"
>
> >but it looks like you have a cascading if/else chain, and so the ENTIRE
> CHAIN is "at line 511".
>
> >Break up the else/ifs and you'll get a better pointer to the offending
> line.

> ------------
>
> Yep, seems like a some line inside of the if/else statement caused the
> error
>
> By the way
>
> if ($action =~ /start/i) {
>         &MainPage($q, \%p);
> } elsif ($action =~ /new|del|edt|save/i ) {
>
> {
>
> seems better to me

It's more compact, but I bet it runs slower.  This might even be a
case where performance matters (i.e. if the cgi script sees a lot of
traffic).


From andy at petdance.com  Wed Mar  5 17:30:21 2008
From: andy at petdance.com (Andy Lester)
Date: Wed, 5 Mar 2008 19:30:21 -0600
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <200803060126.m261QGCR094802@kzsu.stanford.edu>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
	<200803060126.m261QGCR094802@kzsu.stanford.edu>
Message-ID: <1B199114-EC6C-462C-91E3-5845CDDAB253@petdance.com>


On Mar 5, 2008, at 7:26 PM, Joe Brenner wrote:

> It's more compact, but I bet it runs slower.  This might even be a
> case where performance matters (i.e. if the cgi script sees a lot of
> traffic).


No CGI script is going to be bottlenecked by a regex deciding an  
action to take, regardless of how much traffic it gets.

xoa

--
Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance





From extasia at extasia.org  Wed Mar  5 17:33:41 2008
From: extasia at extasia.org (David Alban)
Date: Wed, 5 Mar 2008 17:33:41 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <200803060126.m261QGCR094802@kzsu.stanford.edu>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
	<200803060126.m261QGCR094802@kzsu.stanford.edu>
Message-ID: <4c714a9c0803051733g55c55f26q3ff66af2344ef787@mail.gmail.com>

with respect to:

  /start/i

consider beginning of string and end of string anchors unless you want
also to match:

  restart
  starts
  started

if it weren't for the /i modifier, i'd say just use 'eq'.
actually--you could do:

  lc $action eq 'start'

>  > if ($action =~ /start/i) {
>  >         &MainPage($q, \%p);


-- 
Live in a world of your own, but always welcome visitors.

From sphink at gmail.com  Wed Mar  5 17:33:47 2008
From: sphink at gmail.com (Steve Fink)
Date: Wed, 5 Mar 2008 17:33:47 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <200803060126.m261QGCR094802@kzsu.stanford.edu>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
	<200803060126.m261QGCR094802@kzsu.stanford.edu>
Message-ID: <7d7f2e8c0803051733s141f0f6k6ef0f90ffa3d100b@mail.gmail.com>

>  > By the way
>  >
>  > if ($action =~ /start/i) {
>  >         &MainPage($q, \%p);
>  > } elsif ($action =~ /new|del|edt|save/i ) {
>  >
>  > {
>  >
>  > seems better to me
>
>  It's more compact, but I bet it runs slower.  This might even be a
>  case where performance matters (i.e. if the cgi script sees a lot of
>  traffic).

I'll take that bet. I bet the all-regex version is faster. Especially
on newer versions of Perl that have the trie optimization, but even
without that I'd guess that it's faster. If I weren't so damn lazy,
I'd write a benchmark to test it. But I can't be bothered, because
there's no way that the performance of that line of code is going to
be significant compared to the processing and page generation. (And if
this is a CGI and not mod_perl, the startup time is going to be orders
of magnitude more than all that anyway.)

From doom at kzsu.stanford.edu  Wed Mar  5 17:51:56 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Wed, 05 Mar 2008 17:51:56 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <7d7f2e8c0803051733s141f0f6k6ef0f90ffa3d100b@mail.gmail.com>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
	<200803060126.m261QGCR094802@kzsu.stanford.edu>
	<7d7f2e8c0803051733s141f0f6k6ef0f90ffa3d100b@mail.gmail.com>
Message-ID: <200803060152.m261puJE095277@kzsu.stanford.edu>


Steve Fink <sphink at gmail.com> wrote:
> >  > By the way
> >  >
> >  > if ($action =~ /start/i) {
> >  >         &MainPage($q, \%p);
> >  > } elsif ($action =~ /new|del|edt|save/i ) {
> >  >
> >  > {
> >  >
> >  > seems better to me
> >
> >  It's more compact, but I bet it runs slower.  This might even be a
> >  case where performance matters (i.e. if the cgi script sees a lot of
> >  traffic).
>
> I'll take that bet. I bet the all-regex version is faster. Especially
> on newer versions of Perl that have the trie optimization, but even
> without that I'd guess that it's faster.

And you may already be a winner.

I did a quickie benchmark and it looks like the regexp alternation
version is around 17% faster even in the "worst" cases (running on
perl v5.8.7).  I've seen some regexps with heavy reliance on alternation
that ran pretty slow, but this isn't one of them (and that was
a while back in any case).

Also, you (and Andy Lester) are probably right about it being a silly
thing to worry about -- this session on premature optimization is now
over.


Code and results follow:

my $count = 10000000;

my $result =
timethese($count, {
                   'logical_ors - start'        => 'chooser_mark_1( "start" )',
                   'regexp_alternation - start' => 'chooser_mark_2( "start" )',
                  });

$result =
timethese($count, {
                   'logical_ors - edt'        => 'chooser_mark_1( "edt" )',
                   'regexp_alternation - edt' => 'chooser_mark_2( "edt" )',
                  });


$result =
timethese($count, {
                   'logical_ors - save'        => 'chooser_mark_1( "save" )',
                   'regexp_alternation - save' => 'chooser_mark_2( "save" )',
                  });

sub chooser_mark_1 {
  my $action = shift;
  my $branch;
  if ($action =~ /start/i) {
    $branch = 1;
  } elsif ($action =~ /new/i ||
           $action =~ /del/i ||
           $action =~ /edt/i ||
           $action =~ /save/i ) {
    $branch = 2;
  }
  return $branch;
}


sub chooser_mark_2 {
  my $action = shift;
  my $branch;
  if ($action =~ /start/i) {
    $branch = 1;
  } elsif ($action =~ /new|del|edt|save/i ) {
    $branch = 2;
  }
  return $branch;
}


Results:

Benchmark: timing 10000000 iterations of logical_ors - start, regexp_alternation - start...
logical_ors - start: 18 wallclock secs (16.96 usr +  0.33 sys = 17.29 CPU) @ 578369.00/s (n=10000000)
regexp_alternation - start: 18 wallclock secs (16.83 usr +  0.33 sys = 17.16 CPU) @ 582750.58/s (n=10000000)
Benchmark: timing 10000000 iterations of logical_ors - edt, regexp_alternation - edt...
logical_ors - edt: 22 wallclock secs (21.67 usr +  0.42 sys = 22.09 CPU) @ 452693.53/s (n=10000000)
regexp_alternation - edt: 17 wallclock secs (18.15 usr +  0.37 sys = 18.52 CPU) @ 539956.80/s (n=10000000)
Benchmark: timing 10000000 iterations of logical_ors - save, regexp_alternation - save...
logical_ors - save: 23 wallclock secs (23.53 usr +  0.48 sys = 24.01 CPU) @ 416493.13/s (n=10000000)
regexp_alternation - save: 19 wallclock secs (18.46 usr +  0.36 sys = 18.82 CPU) @ 531349.63/s (n=10000000)





> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

From nheller at silcon.com  Wed Mar  5 19:12:15 2008
From: nheller at silcon.com (Neil Heller)
Date: Wed, 5 Mar 2008 19:12:15 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
Message-ID: <000001c87f37$e15a4ee0$a40eeca0$@com>


The "at line 511" really means "the statement that starts on line 511"  
but it looks like you have a cascading if/else chain, and so the  
ENTIRE CHAIN is "at line 511".

Break up the else/ifs and you'll get a better pointer to the offending  
line.

Thud (sound of the heal of my hand hitting my forehead).

Thanks a lot.  I even think I know which one is the offender:  the one I
wrote.

Neil Heller



From doom at kzsu.stanford.edu  Wed Mar  5 19:51:46 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Wed, 05 Mar 2008 19:51:46 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <200803060152.m261puJE095277@kzsu.stanford.edu>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
	<200803060126.m261QGCR094802@kzsu.stanford.edu>
	<7d7f2e8c0803051733s141f0f6k6ef0f90ffa3d100b@mail.gmail.com>
	<200803060152.m261puJE095277@kzsu.stanford.edu>
Message-ID: <200803060351.m263pkug096743@kzsu.stanford.edu>


Joe Brenner <doom at kzsu.stanford.edu> wrote:
> Steve Fink <sphink at gmail.com> wrote:
> > >  > By the way
> > >  >
> > >  > if ($action =~ /start/i) {
> > >  >         &MainPage($q, \%p);
> > >  > } elsif ($action =~ /new|del|edt|save/i ) {
> > >  >
> > >  > {
> > >  >
> > >  > seems better to me
> > >
> > >  It's more compact, but I bet it runs slower.  This might even be a
> > >  case where performance matters (i.e. if the cgi script sees a lot of
> > >  traffic).
> >
> > I'll take that bet. I bet the all-regex version is faster. Especially
> > on newer versions of Perl that have the trie optimization, but even
> > without that I'd guess that it's faster.
>
> And you may already be a winner.
>
> I did a quickie benchmark and it looks like the regexp alternation
> version is around 17% faster even in the "worst" cases (running on
> perl v5.8.7).  I've seen some regexps with heavy reliance on alternation
> that ran pretty slow, but this isn't one of them (and that was
> a while back in any case).

Funny thing though... just for the hell of it I ran the same code under
perl 5.10 (really, 5.9.5), and the results came out the other way, with
the alternation regexp being a bit slower (~25%) than the stack of
boolean ors.  Weird.

No, I take it back, it's not that weird: the Debian/Ubuntu version of
5.8.7 was built very differently from my 5.9.5 build:

perl 5.8.7:
  Compile-time options: MULTIPLICITY USE_ITHREADS USE_64_BIT_INT
                        USE_64_BIT_ALL USE_LARGE_FILES
                        PERL_IMPLICIT_CONTEXT

perl 5.9.5:
  Compile-time options: PERL_DONT_CREATE_GVSV PERL_MALLOC_WRAP
                        USE_64_BIT_ALL USE_64_BIT_INT USE_LARGE_FILES
                        USE_PERLIO


From biztos at mac.com  Wed Mar  5 21:58:30 2008
From: biztos at mac.com (Kevin Frost)
Date: Wed, 5 Mar 2008 21:58:30 -0800
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <200803060351.m263pkug096743@kzsu.stanford.edu>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
	<200803060126.m261QGCR094802@kzsu.stanford.edu>
	<7d7f2e8c0803051733s141f0f6k6ef0f90ffa3d100b@mail.gmail.com>
	<200803060152.m261puJE095277@kzsu.stanford.edu>
	<200803060351.m263pkug096743@kzsu.stanford.edu>
Message-ID: <1F9AF634-A001-4B5E-BA9E-E63BBA816F59@mac.com>

There is also a Switch module if you miss your case/esac.

Haven't used it in years and never benchmarked it, and now I would  
rather lean towards a dispatch model, but IIRC it did its job well.

  -- frosty

(via iPhone)


On Mar 5, 2008, at 7:51 PM, Joe Brenner <doom at kzsu.stanford.edu> wrote:

>
> Joe Brenner <doom at kzsu.stanford.edu> wrote:
>> Steve Fink <sphink at gmail.com> wrote:
>>>>> By the way
>>>>>
>>>>> if ($action =~ /start/i) {
>>>>>        &MainPage($q, \%p);
>>>>> } elsif ($action =~ /new|del|edt|save/i ) {
>>>>>
>>>>> {
>>>>>
>>>>> seems better to me
>>>>
>>>> It's more compact, but I bet it runs slower.  This might even be a
>>>> case where performance matters (i.e. if the cgi script sees a lot  
>>>> of
>>>> traffic).
>>>
>>> I'll take that bet. I bet the all-regex version is faster.  
>>> Especially
>>> on newer versions of Perl that have the trie optimization, but even
>>> without that I'd guess that it's faster.
>>
>> And you may already be a winner.
>>
>> I did a quickie benchmark and it looks like the regexp alternation
>> version is around 17% faster even in the "worst" cases (running on
>> perl v5.8.7).  I've seen some regexps with heavy reliance on  
>> alternation
>> that ran pretty slow, but this isn't one of them (and that was
>> a while back in any case).
>
> Funny thing though... just for the hell of it I ran the same code  
> under
> perl 5.10 (really, 5.9.5), and the results came out the other way,  
> with
> the alternation regexp being a bit slower (~25%) than the stack of
> boolean ors.  Weird.
>
> No, I take it back, it's not that weird: the Debian/Ubuntu version of
> 5.8.7 was built very differently from my 5.9.5 build:
>
> perl 5.8.7:
>  Compile-time options: MULTIPLICITY USE_ITHREADS USE_64_BIT_INT
>                        USE_64_BIT_ALL USE_LARGE_FILES
>                        PERL_IMPLICIT_CONTEXT
>
> perl 5.9.5:
>  Compile-time options: PERL_DONT_CREATE_GVSV PERL_MALLOC_WRAP
>                        USE_64_BIT_ALL USE_64_BIT_INT USE_LARGE_FILES
>                        USE_PERLIO
>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

From andy at petdance.com  Wed Mar  5 22:03:21 2008
From: andy at petdance.com (Andy Lester)
Date: Thu, 6 Mar 2008 00:03:21 -0600
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <1F9AF634-A001-4B5E-BA9E-E63BBA816F59@mac.com>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<052BC823-ACA7-47B4-A7DD-DB4AA974F521@petdance.com>
	<09AD1BE32C2FCF47A8049BDA1D0449ED0242CF8F@Site1Mail04.jeeves.ask.info>
	<200803060126.m261QGCR094802@kzsu.stanford.edu>
	<7d7f2e8c0803051733s141f0f6k6ef0f90ffa3d100b@mail.gmail.com>
	<200803060152.m261puJE095277@kzsu.stanford.edu>
	<200803060351.m263pkug096743@kzsu.stanford.edu>
	<1F9AF634-A001-4B5E-BA9E-E63BBA816F59@mac.com>
Message-ID: <4A0F0920-A72D-4D48-A892-63BDE41DA43C@petdance.com>


On Mar 5, 2008, at 11:58 PM, Kevin Frost wrote:

> There is also a Switch module if you miss your case/esac.


And Perl 5.10 gives us given/when.

xoa

--
Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance





From nheller at silcon.com  Sat Mar  8 10:19:29 2008
From: nheller at silcon.com (Neil Heller)
Date: Sat, 8 Mar 2008 10:19:29 -0800
Subject: [sf-perl] Debugging a CGI script - how
Message-ID: <000001c88148$f32cff60$d986fe20$@com>

I've been modifying a CGI script and have run into a problem that bring to
mind a larger question.

My script uses data from a relational database to pull information about
external files.  The script then displays information from the relational
database and the contents of the external file.  The script then continues
on the next iteration - each iteration equals one entry in the relational
database.

If there are 3 or fewer entries in the relational database, the script works
well.  At the fourth entry in the relational database, Apache throws an
error and the user (me at this point) receives only a generic-type error.

Does anybody have any ideas about how to best go about finding the error?

Neil Heller


From david at fetter.org  Sat Mar  8 11:17:33 2008
From: david at fetter.org (David Fetter)
Date: Sat, 8 Mar 2008 11:17:33 -0800
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <000001c88148$f32cff60$d986fe20$@com>
References: <000001c88148$f32cff60$d986fe20$@com>
Message-ID: <20080308191733.GG23803@fetter.org>

On Sat, Mar 08, 2008 at 10:19:29AM -0800, Neil Heller wrote:
> I've been modifying a CGI script and have run into a problem that
> bring to mind a larger question.
> 
> My script uses data from a relational database to pull information
> about external files.  The script then displays information from the
> relational database and the contents of the external file.  The
> script then continues on the next iteration - each iteration equals
> one entry in the relational database.
> 
> If there are 3 or fewer entries in the relational database, the
> script works well.  At the fourth entry in the relational database,
> Apache throws an error and the user (me at this point) receives only
> a generic-type error.
> 
> Does anybody have any ideas about how to best go about finding the
> error?

Does anybody have a reference on how to write a good bug report/shout
out for help?  I've seen Eric Raymond's smarmy and condescending "How
to Ask Smart Questions," but it misses important points, and even if
it didn't, a person reading it is unlikely to get further than the
first couple of paragraphs unless that person is a core-deep
masochist.

Anybody?

Cheers,
David.
-- 
David Fetter <david at fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter at gmail.com

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

From doom at kzsu.stanford.edu  Sat Mar  8 12:32:02 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Sat, 08 Mar 2008 12:32:02 -0800
Subject: [sf-perl] Debugging a CGI script - how
In-Reply-To: <000001c88148$f32cff60$d986fe20$@com>
References: <000001c88148$f32cff60$d986fe20$@com>
Message-ID: <200803082032.m28KW2Iu048334@kzsu.stanford.edu>


Neil Heller <nheller at silcon.com> wrote:

> I've been modifying a CGI script and have run into a problem that bring to
> mind a larger question.
>
> My script uses data from a relational database to pull information about
> external files.  The script then displays information from the relational
> database and the contents of the external file.  The script then continues
> on the next iteration - each iteration equals one entry in the relational
> database.
>
> If there are 3 or fewer entries in the relational database, the script works
> well.  At the fourth entry in the relational database, Apache throws an
> error and the user (me at this point) receives only a generic-type error.
>
> Does anybody have any ideas about how to best go about finding the error?

Have you checked the apache error log?  Where are you seeing this
"generic error message"?

Note that if you toss in statements like:

  print STDERR "The value is: $value at " . __LINE__ . "\n";

That output should show up in the apache error log.

At a guess, you should look at the data involved more closely: I would
guess there's something in your fourth record that's messing you up,
and you need to be more careful with quoting or using bind parameter
("?" placeholders) or something like that.

Try some runs with different data, organized in different ways.

I also might suggest that at some point you shoule spend some time to
learn how to use the perl debugger, but that admittedly can be a bit
of a pain in a web context.

You might try to cut down your code the smallest example that reproduces
the problem, then you could post the code here, and we'll look it
over.


From jared at wordzoo.com  Sat Mar  8 12:37:10 2008
From: jared at wordzoo.com (Jared Rhine)
Date: Sat, 8 Mar 2008 12:37:10 -0800
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080308191733.GG23803@fetter.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
Message-ID: <BF83088F-3199-4787-BA79-1DFF3786DA35@wordzoo.com>

> Does anybody have a reference on how to write a good bug report/shout
> out for help?

I've no bug-reporting guides that aren't application-specific.  It  
seems that the crankiness experienced on this list and others boils  
down to these two rules:

1. Cut-and-paste the actual error message, not your summary of the  
error message
2. Include actual code

Time and time again, inquires which include at least one of these  
pieces of information get an answer.  Inquiries which skip both are  
barely worth responding to, except to repeat the above two rules.

For Perl-list questions, Raymond's document should be shortened  
dramatically by basing around these two.  (He was trying to construct  
a general method of question-asking, which is inherently harder than  
any procedure for a specific context, like an app or Perl coding.)

("generic-type error" is the funniest problem report I've seen in a  
while, FWIW.  Probably what's going on here is the original poster is  
seeing a 500 error in Apache, which is indeed a generic-type error  
with no script-relevant info.  They haven't learned yet to look in the  
error logs.)

For what it's worth, I can see how beginners are so unstructured in  
their thinking as to attribute a sort of magical ability to more  
experienced people, to the point where their mental model is that  
senior developers know what the problem is from just a hand-wavey  
description of the problem.  It's half that people don't know how to  
Ask Good Questions, but also that they don't understand the process  
experienced people use to solve problems.  They can't distinguish  
between problems which have one or two possible answers and problems  
which have 1000 possible answers.

When I train junior staff, it's amazing how the most effective lesson  
I can teach is generally "read the actual error message".  "No, read  
it".  "Repeat each word to me".  "It's telling you the exact  
problem".  "Rephrase the error message to me".  Of course, experienced  
people know error messages often lie or are useless, but errors  
received by junior staff usually aren't obscure but very plain.

-- Jared


From doom at kzsu.stanford.edu  Sat Mar  8 12:38:04 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Sat, 08 Mar 2008 12:38:04 -0800
Subject: [sf-perl] Getting Help (was Re: Debugging a CGI script - how)
In-Reply-To: <20080308191733.GG23803@fetter.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
Message-ID: <200803082038.m28Kc44f048436@kzsu.stanford.edu>


David Fetter <david at fetter.org> wrote:
> Neil Heller wrote:

> > I've been modifying a CGI script and have run into a problem that
> > bring to mind a larger question.
> >
> > My script uses data from a relational database to pull information
> > about external files.  The script then displays information from the
> > relational database and the contents of the external file.  The
> > script then continues on the next iteration - each iteration equals
> > one entry in the relational database.
> >
> > If there are 3 or fewer entries in the relational database, the
> > script works well.  At the fourth entry in the relational database,
> > Apache throws an error and the user (me at this point) receives only
> > a generic-type error.
> >
> > Does anybody have any ideas about how to best go about finding the
> > error?
>
> Does anybody have a reference on how to write a good bug report/shout
> out for help?

I know the problem you're talking about, but Neil's question doesn't
seem *all* that bad to me.  It's a pretty generic question about how to
debug weird problems in web/db programming...

> I've seen Eric Raymond's smarmy and condescending "How to Ask Smart
> Questions," but it misses important points, and even if it didn't, a
> person reading it is unlikely to get further than the first couple of
> paragraphs unless that person is a core-deep masochist.

At the very least, it's certainly too verbose.  (And as I remember it,
Rick Moen has essentially disowned it at this point.)

> Anybody?

Some projects write their own.  There are a few pages on the subject
written for the emacs documentation:

  http://www.gnu.org/software/emacs/manual/html_node/emacs/Bugs.html


From doom at kzsu.stanford.edu  Sat Mar  8 12:43:11 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Sat, 08 Mar 2008 12:43:11 -0800
Subject: [sf-perl] Getting Help (was Re: Debugging a CGI script - how)
In-Reply-To: <200803082038.m28Kc44f048436@kzsu.stanford.edu>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<200803082038.m28Kc44f048436@kzsu.stanford.edu>
Message-ID: <200803082043.m28KhBxY048532@kzsu.stanford.edu>


Joe Brenner <doom at kzsu.stanford.edu> wrote:

> Some projects write their own.  There are a few pages on the subject
> written for the emacs documentation:
>
>   http://www.gnu.org/software/emacs/manual/html_node/emacs/Bugs.html

Ah, of course.  There's Mark Jason Dominus on the subject:

   http://perl.plover.com/Questions.html


From biztos at mac.com  Sat Mar  8 12:54:11 2008
From: biztos at mac.com (Kevin Frost)
Date: Sat, 8 Mar 2008 12:54:11 -0800
Subject: [sf-perl] Debugging a CGI script - how
In-Reply-To: <200803082032.m28KW2Iu048334@kzsu.stanford.edu>
References: <000001c88148$f32cff60$d986fe20$@com>
	<200803082032.m28KW2Iu048334@kzsu.stanford.edu>
Message-ID: <40389DDC-69A7-4F27-9116-7B74E23A690D@mac.com>

I would also add that if you're doing fetchrow_arrayref or similar you  
might want to temporarily switch to _hashref and use the Dumper:

use Data::Dumper qw(Dumper);

# ...set up db connection and query, etc.


my $rownum = 0;
while (my $row  = $sth->fetchrow_hashref) {
	# 4th on is tricky...
	$rownum ++;
	printf STDERR "ROW %d: %s\n", $rownum, Dumper($row);

}

...or words to that effect.

-- f.

On Mar 8, 2008, at 12:32 PM, Joe Brenner wrote:

>
>  print STDERR "The value is: $value at " . __LINE__ . "\n";


From doom at kzsu.stanford.edu  Sat Mar  8 13:19:38 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Sat, 08 Mar 2008 13:19:38 -0800
Subject: [sf-perl] Debugging a CGI script - how
In-Reply-To: <40389DDC-69A7-4F27-9116-7B74E23A690D@mac.com>
References: <000001c88148$f32cff60$d986fe20$@com>
	<200803082032.m28KW2Iu048334@kzsu.stanford.edu>
	<40389DDC-69A7-4F27-9116-7B74E23A690D@mac.com>
Message-ID: <200803082119.m28LJcnW049109@kzsu.stanford.edu>


Kevin Frost <biztos at mac.com> wrote:

> Joe Brenner wrote:
> >  print STDERR "The value is: $value at " . __LINE__ . "\n";

> I would also add that if you're doing fetchrow_arrayref or similar you
> might want to temporarily switch to _hashref and use the Dumper:
>
> use Data::Dumper qw(Dumper);
>
> # ...set up db connection and query, etc.
>
>
> my $rownum = 0;
> while (my $row  = $sth->fetchrow_hashref) {
> 	# 4th on is tricky...
> 	$rownum ++;
> 	printf STDERR "ROW %d: %s\n", $rownum, Dumper($row);
>
> }
>
> ...or words to that effect.

Sure, sounds good.

I just thought I'd make the point that there isn't any particularly need
to avoid using fetchrow_hashref.  As I remember it, the performance
penalty on doing that was fixed a while back.


From garth.webb at gmail.com  Sat Mar  8 13:21:34 2008
From: garth.webb at gmail.com (Garth Webb)
Date: Sat, 8 Mar 2008 13:21:34 -0800
Subject: [sf-perl] Debugging a CGI script - how
In-Reply-To: <200803082032.m28KW2Iu048334@kzsu.stanford.edu>
References: <000001c88148$f32cff60$d986fe20$@com>
	<200803082032.m28KW2Iu048334@kzsu.stanford.edu>
Message-ID: <c0d982e10803081321j6b4b9710he174f7daec83e420@mail.gmail.com>

On Sat, Mar 8, 2008 at 12:32 PM, Joe Brenner <doom at kzsu.stanford.edu> wrote:
>
>  Neil Heller <nheller at silcon.com> wrote:
>
>  > I've been modifying a CGI script and have run into a problem that bring to
>  > mind a larger question.
>  >
>  > My script uses data from a relational database to pull information about
>  > external files.  The script then displays information from the relational
>  > database and the contents of the external file.  The script then continues
>  > on the next iteration - each iteration equals one entry in the relational
>  > database.
>  >
>  > If there are 3 or fewer entries in the relational database, the script works
>  > well.  At the fourth entry in the relational database, Apache throws an
>  > error and the user (me at this point) receives only a generic-type error.
>  >
>  > Does anybody have any ideas about how to best go about finding the error?
>
>  Have you checked the apache error log?  Where are you seeing this
>  "generic error message"?
>
>  Note that if you toss in statements like:
>
>   print STDERR "The value is: $value at " . __LINE__ . "\n";
>
>  That output should show up in the apache error log.
>
>  At a guess, you should look at the data involved more closely: I would
>  guess there's something in your fourth record that's messing you up,
>  and you need to be more careful with quoting or using bind parameter
>  ("?" placeholders) or something like that.
>
>  Try some runs with different data, organized in different ways.
>
>  I also might suggest that at some point you shoule spend some time to
>  learn how to use the perl debugger, but that admittedly can be a bit
>  of a pain in a web context.

I would second learning the Perl debugger.  Its invaluable for quickly
getting to the root cause of your problems.  I don't think being in a
web context really changes much.  If its a straight CGI script (not
mod_perl) then you can easily run the script in the debugger and
supply your own parameters:

  $ perl -d your_script.cgi
  DB<1> CGI::param('foo' => 'bar');

Then later in your code, this will work how you expect:

  my $foo = CGI::param('foo')

If you put all your functionality into separate methods, its even
easier because you can just call the method with the args you need,
and you don't even have to run the whole program:

  DB<2> s my_method({arg1 => 'someval', arg2 => [1, 2, 3])

where 's' means 'step into'.  If you really want/need to run the
script directly from a real web request, its possible to start up
apache with the perl debugger running and step through requests as
they happen.

If your using mod_perl, you can still use the debugger, its just a
little more complicated if you want to debug calls that involve
mod_perl itself.  Most of the time though, I don't really need to
debug that code.

At work we run mod_perl and all our code is in separate Perl modules
that I can debug easily and individually outside mod_perl:

  $ perl -d -e 1
  DB<1> use Some::Data::Class
  DB<2> $res = Some::Data::Class->load_results(type => 'example')
  DB<3> x $res

which starts up the perl debugger in an interactive mode, loads a
module, makes a call to a method and then eXamines the results.

Garth

>  You might try to cut down your code the smallest example that reproduces
>  the problem, then you could post the code here, and we'll look it
>  over.
>
>
>
>  _______________________________________________
>  SanFrancisco-pm mailing list
>  SanFrancisco-pm at pm.org
>  http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>

From quinn at fairpath.com  Sat Mar  8 13:29:15 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sat, 8 Mar 2008 13:29:15 -0800
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080308191733.GG23803@fetter.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
Message-ID: <20080308212915.GA62053@fu.funkspiel.org>

On Sat, Mar 08, 2008 at 11:17:33AM -0800, David Fetter wrote:

> Does anybody have a reference on how to write a good bug report/shout
> out for help?

Unfortunately, I don't.  I'd be interested to know of one and, if
someone can find one that looks good enough, I will include it in the
metadata (intro email, et cetera) for this list.

Meanwhile, for the purposes of this conversation:

Neil, we need a lot more detail and code samples if we're going to debug
this specific problem.

On the other hand, if you just want general tips about debugging
Apache scripts (so you can pursue it on your own), Joe's recommendations
are spot-on.

I'll add that it's possible to connect the Perl debugger to an Apache
process.  Matt Lanier gave a nice talk on this some years ago.
Here are his notes:
http://www.lanier.org/~matt/ad-talk/launch-perldb.html

Meanwhile, since you're just starting out with Apache and have lots of
questions, you should pick up one of the books on the subject.
If you're using Apache 1.3.x (as opposed to Apache 2), I highly recommend the
mod_perl Developer's Cookbook.  It's very well-written, and it's
broken up into "recipes" so you can get at exactly what you want
(they're self-contained, so that you need little or no prior knowledge
to benefit from a given recipe).  You can peelk at the table of contents
and some samples here:  http://www.modperlcookbook.org/

Thanks,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From quinn at fairpath.com  Sat Mar  8 13:44:32 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sat, 8 Mar 2008 13:44:32 -0800
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080308212915.GA62053@fu.funkspiel.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080308212915.GA62053@fu.funkspiel.org>
Message-ID: <20080308214432.GA62111@fu.funkspiel.org>

On Sat, Mar 08, 2008 at 01:29:15PM -0800, Quinn Weaver wrote:
> On Sat, Mar 08, 2008 at 11:17:33AM -0800, David Fetter wrote:
> 
> > Does anybody have a reference on how to write a good bug report/shout
> > out for help?
> 
> Unfortunately, I don't.  I'd be interested to know of one and, if
> someone can find one that looks good enough, I will include it in the
> metadata (intro email, et cetera) for this list.

Actually, I decided to write one myself.  Here's a first stab.  Input
is welcome.  (I don't promise I'll take all advice, but I'll at least
listen. :) )

*******************

If you have some problem code and want to ask a question about it, please:

- First try a quick web search (Google or PerlMonks), and consult
  the standard docs bundled with Perl.  It's amazing how many questions
  are resolved this way.

If you do post, please:

- Be polite.

- Include relevant code snippets, input data, output, and configuration files.
  Send the smallest bit that you think is relevant, not the whole thing.

- Describe the circumstances in which the problem appears, what you've
  already done to try to diagnose and fix it, and why you're stuck.

- Be patient with the speed at which people respond.  After all, these are
  experts who are giving you a free service!

Thanks,
The San Francisco Perl Mongers staff

*******************

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From quinn at fairpath.com  Sat Mar  8 13:46:28 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sat, 8 Mar 2008 13:46:28 -0800
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080308212915.GA62053@fu.funkspiel.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080308212915.GA62053@fu.funkspiel.org>
Message-ID: <20080308214628.GA62147@fu.funkspiel.org>

On Sat, Mar 08, 2008 at 01:29:15PM -0800, Quinn Weaver wrote:

> I'll add that it's possible to connect the Perl debugger to an Apache
> process.  Matt Lanier gave a nice talk on this some years ago.
> Here are his notes:
> http://www.lanier.org/~matt/ad-talk/launch-perldb.html

Clarification:  this is for mod_perl.  For CGI scripts, follow Garth's advice
instead.

Thanks,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From doom at kzsu.stanford.edu  Sat Mar  8 13:50:30 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Sat, 08 Mar 2008 13:50:30 -0800
Subject: [sf-perl] Getting Help (was Re: Debugging a CGI script - how)
In-Reply-To: <200803082043.m28KhBxY048532@kzsu.stanford.edu>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<200803082038.m28Kc44f048436@kzsu.stanford.edu>
	<200803082043.m28KhBxY048532@kzsu.stanford.edu>
Message-ID: <200803082150.m28LoUor049603@kzsu.stanford.edu>


Joe Brenner <doom at kzsu.stanford.edu> wrote:
> Joe Brenner <doom at kzsu.stanford.edu> wrote:
>
> > Some projects write their own.  There are a few pages on the subject
> > written for the emacs documentation:
> >
> >   http://www.gnu.org/software/emacs/manual/html_node/emacs/Bugs.html
>
> Ah, of course.  There's Mark Jason Dominus on the subject:
>
>    http://perl.plover.com/Questions.html

And by an amazing coincidence, look what just turned up on the new
"official" perl wiki:

   http://www.perlfoundation.org/perl5/index.cgi?how_to_ask_perl_questions


From quinn at fairpath.com  Sat Mar  8 14:08:57 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sat, 8 Mar 2008 14:08:57 -0800
Subject: [sf-perl] Getting Help (was Re: Debugging a CGI script - how)
In-Reply-To: <200803082150.m28LoUor049603@kzsu.stanford.edu>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<200803082038.m28Kc44f048436@kzsu.stanford.edu>
	<200803082043.m28KhBxY048532@kzsu.stanford.edu>
	<200803082150.m28LoUor049603@kzsu.stanford.edu>
Message-ID: <20080308220857.GA62339@fu.funkspiel.org>

On Sat, Mar 08, 2008 at 01:50:30PM -0800, Joe Brenner wrote:
> 
> Joe Brenner <doom at kzsu.stanford.edu> wrote:
> > Joe Brenner <doom at kzsu.stanford.edu> wrote:
> >
> > > Some projects write their own.  There are a few pages on the subject
> > > written for the emacs documentation:
> > >
> > >   http://www.gnu.org/software/emacs/manual/html_node/emacs/Bugs.html
> >
> > Ah, of course.  There's Mark Jason Dominus on the subject:
> >
> >    http://perl.plover.com/Questions.html
> 
> And by an amazing coincidence, look what just turned up on the new
> "official" perl wiki:
> 
>    http://www.perlfoundation.org/perl5/index.cgi?how_to_ask_perl_questions

Nice.  Maybe I'll just use that.

Meanwhile, here's a stab at the other side of things:

*****************

How to answer questions:

- Be polite.  If someone is rude or out of line, correct them firmly
  but civilly.

- Wherever possible, give people info that empowers them, so they can
  solve problems on their own.

- If you need more information, ask for it.  No need to guess what the
  poster is thinking.

- Remember you were once a newbie too.

*****************

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From extasia at extasia.org  Sat Mar  8 14:09:46 2008
From: extasia at extasia.org (David Alban)
Date: Sat, 8 Mar 2008 14:09:46 -0800
Subject: [sf-perl] Getting Help (was Re: Debugging a CGI script - how)
In-Reply-To: <20080308214432.GA62111@fu.funkspiel.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080308212915.GA62053@fu.funkspiel.org>
	<20080308214432.GA62111@fu.funkspiel.org>
Message-ID: <4c714a9c0803081409w5ad72735ta02ed194210b36a7@mail.gmail.com>

there's overlap here with what joe wrote below, but also consider...

  * demonstrate a serious effort on your part before posting your
questions (see 'rtfm'); at the very least, have done a web search and
have consulted the available documentation (see 'perldoc', 'man', and
friends)

  * specific questions will get much better results than extremely
high level / vague / general questions or questions without detail

  * don't post anything that even appears that you want someone else
to do your work for free; note the use of the word "appears"; you may
not intend it, but if it appears that you do you will probably not get
the information you need

  * try and come up with the smallest and simplest code that
demonstrates your problem (if applicable), you'll get fewer responses
if folks have to spend more then five or ten minutes grokking your
problem.

  * if someone cranky flames you, don't respond to the flamey bits.  just don't.

On Sat, Mar 8, 2008 at 1:44 PM, Quinn Weaver <quinn at fairpath.com> wrote:
>  If you have some problem code and want to ask a question about it, please:
>
>  - First try a quick web search (Google or PerlMonks), and consult
>   the standard docs bundled with Perl.  It's amazing how many questions
>   are resolved this way.
>
>  If you do post, please:
>
>  - Be polite.
>
>  - Include relevant code snippets, input data, output, and configuration files.
>   Send the smallest bit that you think is relevant, not the whole thing.
>
>  - Describe the circumstances in which the problem appears, what you've
>   already done to try to diagnose and fix it, and why you're stuck.
>
>  - Be patient with the speed at which people respond.  After all, these are
>   experts who are giving you a free service!
>
>  Thanks,
>  The San Francisco Perl Mongers staff

-- 
Live in a world of your own, but always welcome visitors.

From quinn at fairpath.com  Sat Mar  8 14:30:43 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Sat, 8 Mar 2008 14:30:43 -0800
Subject: [sf-perl] Getting Help (was Re: Debugging a CGI script - how)
In-Reply-To: <4c714a9c0803081409w5ad72735ta02ed194210b36a7@mail.gmail.com>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080308212915.GA62053@fu.funkspiel.org>
	<20080308214432.GA62111@fu.funkspiel.org>
	<4c714a9c0803081409w5ad72735ta02ed194210b36a7@mail.gmail.com>
Message-ID: <20080308223043.GA62452@fu.funkspiel.org>

On Sat, Mar 08, 2008 at 02:09:46PM -0800, David Alban wrote:
> there's overlap here with what joe wrote below, but also consider...

This is great.  Keep the good ideas coming, everyone,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From rick at linuxmafia.com  Mon Mar 10 03:49:08 2008
From: rick at linuxmafia.com (Rick Moen)
Date: Mon, 10 Mar 2008 03:49:08 -0700
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080308191733.GG23803@fetter.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
Message-ID: <20080310104908.GC3420@linuxmafia.com>

Quoting David Fetter (david at fetter.org):

> I've seen Eric Raymond's smarmy and condescending "How
> to Ask Smart Questions"....

He had some insufferably smarmy and condescending co-author, too.
(Probably a Scandinavian, judging by the guy's surname:  My Tante Bjorg
always warned me to beware of them.)

That essay has become much too verbose and oddly organised:  As Tolkien
put it, it grew in the telling.  I'm aiming at some point to write
something a whole lot shorter and punchier, with some sort of much more
link-ey and less linear structure.

I really rather like the Java Ranch variant, FWIW:
http://faq.javaranch.com/java/HowToAskQuestionsOnJavaRanch

-- 
Cheers,                                      "Reality is not optional."
Rick Moen                                             -- Thomas Sowell
rick at linuxmafia.com

From david at fetter.org  Mon Mar 10 09:03:03 2008
From: david at fetter.org (David Fetter)
Date: Mon, 10 Mar 2008 09:03:03 -0700
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080310104908.GC3420@linuxmafia.com>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080310104908.GC3420@linuxmafia.com>
Message-ID: <20080310160303.GN23803@fetter.org>

On Mon, Mar 10, 2008 at 03:49:08AM -0700, Rick Moen wrote:
> Quoting David Fetter (david at fetter.org):
> 
> > I've seen Eric Raymond's smarmy and condescending "How to Ask
> > Smart Questions"....
> 
> He had some insufferably smarmy and condescending co-author, too.

You missed a great chance to let Mr. Raymond take the blame for this
one.  Very sad, really.

> That essay has become much too verbose and oddly organised:

Its verbosity is not even in the same league of problem as its
authorial voice, which is quite often mistaken for that of a spoiled
14-year-old boy.

> As Tolkien put it, it grew in the telling.  I'm aiming at some point
> to write something a whole lot shorter and punchier, with some sort
> of much more link-ey and less linear structure.
> 
> I really rather like the Java Ranch variant, FWIW:
> http://faq.javaranch.com/java/HowToAskQuestionsOnJavaRanch

The voice of this one (as of this writing) is just great. :)

Omnibus thing for potential question askers to consider:
http://wooledge.org/mywiki/XyProblem

Cheers,
David.
-- 
David Fetter <david at fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter at gmail.com

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate

From andy at petdance.com  Mon Mar 10 09:06:45 2008
From: andy at petdance.com (Andy Lester)
Date: Mon, 10 Mar 2008 11:06:45 -0500
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080310160303.GN23803@fetter.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080310104908.GC3420@linuxmafia.com>
	<20080310160303.GN23803@fetter.org>
Message-ID: <F8205E14-FEE2-43B1-8C3A-C03DDFA9173E@petdance.com>


On Mar 10, 2008, at 11:03 AM, David Fetter wrote:

>> I really rather like the Java Ranch variant, FWIW:
>> http://faq.javaranch.com/java/HowToAskQuestionsOnJavaRanch
>
> The voice of this one (as of this writing) is just great. :)


That page claims "Do not send Java questions to the authors of these  
documents! If you do, they will flame you and you will deserve it."   
That's not correct.  No one deserves to be flamed.

xoxo,
Andy

--
Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance





From rick at linuxmafia.com  Mon Mar 10 14:48:51 2008
From: rick at linuxmafia.com (Rick Moen)
Date: Mon, 10 Mar 2008 14:48:51 -0700
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080310160303.GN23803@fetter.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080310104908.GC3420@linuxmafia.com>
	<20080310160303.GN23803@fetter.org>
Message-ID: <20080310214851.GD3420@linuxmafia.com>

Quoting David Fetter (david at fetter.org):

> You missed a great chance to let Mr. Raymond take the blame for this
> one.  Very sad, really.

Hmm, I though you knew me better than that, David.  I not only never
evade responsibility for what I write (or co-write); I insist on it.  

Particular passages in "How to Ask Questions the Smart Way" still evince
very much my own authorial "voice", one quite different from Eric's.  In
those passages, I aimed to tell readers some blunt, stark, and highly
useful, though widely resented and unpalatable, truths.  Why?  Because I
felt my target audience deserved to hear relevant truths as best I could
see and tell them.  

My point?  I suspect you'd cordially dislike -- resent -- the parts *I*
wrote, too.  Oh well; good  thing I wasn't trying to win any popularity
contests, just tell my intended audience what they needed to know.
Here's an example:

   Describe the problem's symptoms, not your guesses

   It's not useful to say what you think is causing your problem.
   (If your diagnostic theories were such hot stuff, would you be
   consulting others for help?)  So, make sure you're listing the raw
   symptoms of what goes wrong, rather than your interpretations and
   theories.  Let your helpers do the interpretation and diagnosis.

[snip examples of good and bad descriptions]

   Since the preceding point seems to be a tough one for many people to
   grasp, here's a phrase to remind you:  "All diagnosticians are from
   Missouri."  That US state's official motto is "Show me" (earned in
   1899, when Congressman Willard D. Vandiver said "I come from a
   country that raises corn and cotton and cockleburs and Democrats, and
   frothy eloquence neither convinces nor satisfies me.  I'm from
   Missouri.  You've got to show me.")  In diagnosticians' case, it's not
   a matter of skepticism, but rather a literal, functional need to see
   whatever is as close as possible to the same raw evidence that you
   see, rather than your surmises and summaries.  Show us.

I've backed out Eric's reference to "hackers", which was his patch:  My
own target audience, when I started my own version of the essay (but
then found out that Eric was attempting approximately the same essay,
and we joined forces) was technical users who seem to wish to
understand, to learn, be competent, and _maybe teach others_ later on --
as opposed to merely wanting a free helpdesk.  Which in part repays my
own debt, to earlier people who tried to give me clue.

Nothing against people willing to serve as free helpdesks for endless
poorly thought out questions from net.random querents who are
fundamentally unwilling to learn.  I admire the former's saintliness,
and hope they have endurance to match their benevolence. 



From quinn at fairpath.com  Mon Mar 10 16:11:53 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Mon, 10 Mar 2008 15:11:53 -0800
Subject: [sf-perl] Getting Help (was Re:  Debugging a CGI script - how)
In-Reply-To: <20080310214851.GD3420@linuxmafia.com>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080310104908.GC3420@linuxmafia.com>
	<20080310160303.GN23803@fetter.org>
	<20080310214851.GD3420@linuxmafia.com>
Message-ID: <20080310231153.GA83912@fu.funkspiel.org>

Hi, guys, list mom here.  Just a gentle pre-emptive request:  can we
please keep this thread on etiquette from turning into an actual
argument? :) It's certainly not a flame war yet, but my Jedi senses
detect some potential rumors of war.

Please keep your cool, but please also consider when to just let the
matter go.  Every time there's a long thread about who is right,
several people unsubscribe for good.

Thanks,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From doom at kzsu.stanford.edu  Mon Mar 10 23:52:46 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Mon, 10 Mar 2008 23:52:46 -0700
Subject: [sf-perl] Getting Help (was Re: Debugging a CGI script - how)
In-Reply-To: <20080310231153.GA83912@fu.funkspiel.org>
References: <000001c88148$f32cff60$d986fe20$@com>
	<20080308191733.GG23803@fetter.org>
	<20080310104908.GC3420@linuxmafia.com>
	<20080310160303.GN23803@fetter.org>
	<20080310214851.GD3420@linuxmafia.com>
	<20080310231153.GA83912@fu.funkspiel.org>
Message-ID: <200803110652.m2B6qkT0090638@kzsu.stanford.edu>


Quinn Weaver <quinn at fairpath.com> wrote:

> Hi, guys, list mom here.  Just a gentle pre-emptive request:  can we
> please keep this thread on etiquette from turning into an actual
> argument? :)

You may have heard me coming.

I wanted to point out that the "Be polite" rule is a local rule,
not necessarily a universal principle, and other forums have the
right to choose differently.  The other approach is to focus on
the other side: "Grow some skin, try to rise above insult, listen
for what's useful to you, and skip what isn't."


From quinn at fairpath.com  Tue Mar 11 12:32:24 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Tue, 11 Mar 2008 11:32:24 -0800
Subject: [sf-perl] Lugradio conference at the Metreon
Message-ID: <20080311193224.GA94624@fu.funkspiel.org>

Wow.  LugRadio is putting on a Linux/open-source conference April 12 and 13
at the Metreon.  Registration is $10 - Google and Dice are picking up
most of the tab.  They have an impressive list of speakers.

It's funny, just yesterday I was mailing Fred about LinuxWorld
preparation (I want to get started now; it's in August).

One month is a pretty tight schedule, but we could pull off a booth
for LugRadio iff we get enough boothies.  The conference is on a
weekend--is anyone up to helping?  I need as many people as possible
to run four-hour shifts; we need two people at the booth at any
given time.  So that's probably eight slots to fill, total.
Call it ten, in case someone needs to cancel at the last minute.

Are people up for this?  I don't want to be half-assed about this, but
if we can get enough people I'm all for it.

----- Forwarded message from George Woolley <george at metaart.org> -----

From: George Woolley <george at metaart.org>
Date: Mon, 10 Mar 2008 19:04:29 -0800
To: Oakland Perl Mongers <oakland at pm.org>
Subject: [oak perl] Fwd: Perl at LugRadio Live USA 2008 in San Francisco



----------  Forwarded Message  ----------

Subject: Perl at LugRadio Live USA 2008 in San Francisco
Date: Monday 10 March 2008 17:10
From: LugRadio <show at lugradio.org>
...

Hi,

I just wanted to drop you a quick note about a community organised and
run Open Source and technology event that we have been working on that
your group may be interested in. It is called LugRadio Live USA 2008
and has over 35 well-known speakers, nearly 40 exhibitors and takes
place on the 12th and 13th April 2008 in San Francisco. It is very
much a community event, and designed for the community - we have been
working on it for the last six months or so with a large number of
volunteers. To make it as accessible as possible for people, we don't
believe in charging the $600-$1200 that other commercial shows do -
our show is just $10 for the full weekend.

We figured your group may be interested in LugRadio Live USA 2008, and
would love to see you there (we also have exhibition space available
if you would like to exhibit your group - it is free to exhibit!).

Anyway, below is the announcement of the show, if you could pass it
onto your members, that would be great. :)

--------

LugRadio Live USA 2008
12th and 13th April 2008
The Metreon, 101 4th St., San Francisco, CA, 94103

Registration for LugRadio Live USA 2008 is now open at
http://lugradio.org/live/USA2008/register, and tickets for the two-day
spectacular cost only $10 for the full weekend, including full access
to all talks, the exhibition, evening events, and a free bag o'swag
for visitors. Pre-registrations also enjoy additional benefits at the
show and the first 50 registrations will receive a free copy of Postal
2: Share The Pain (subject to age verification), thanks to Running
With Scissors. All pre-registrations will also be entered into a
raffle to win prizes by a variety of vendors.

LugRadio Live USA 2008, the 'rock-conference' from the team behind the
popular LugRadio podcast (http://www.lugradio.org), brings the
successful and unique formula of the UK LugRadio Live events to The
Metreon in San Francisco on the 12th and 13th April 2008. The event is
supported extensively by Google and also supported by Dice.
LinuxQuestions.org is a press partner.

LugRadio Live USA 2008 brings together over 35 speakers across three
stages, 30+ exhibitors, a range of BOF sessions, debate panels,
lightbulb talk sessions, demos and much more, all wrapped up in the
unique event that the UK incarnation has become known for, combining
an incredibly loose, social, inclusive, and amusing atmosphere - if
you are new to LugRadio Live, it is nothing you will have seen before.

Confirmed speakers for the event include:

 * Miguel de Icaza (Mono / Novell / Co-Founder Of GNOME)
 * Ian Murdoch (OpenSolaris / Founder Of Debian)
 * Robert Love (GNOME / Google)
 * Aza Raskin (Mozilla / Humanized)
 * Benjamin Mako Hill (Ubuntu / Debian / FSF)
 * John Buckman (Magnatune)
 * Val Henson (Kernel / VAH consulting)
 * Christopher Blizzard (Mozilla / GNOME)
 * Mike Linksvayer (Creative Commons)
 * Alex Graveley (VMWare / GNOME)
 * David Schleef (GStreamer)
 * Matthew Garrett (Power Management / Kernel)
 * Danese Cooper (Intel / OSI)
 * Aaron Bockover (Banshee / Novell)
 * Liana Holmberg (Second Life / Linden Lab)
 * Emma Jane Hogbin (Hick Tech)
 * Joe Zonker Brockmeier (OpenSuSE / Novell)
 * Kristen Accardi (Kernel)
 * David Lefty Schlesinger
 * Joe Born (Neuros)
 * Selena Deckelmann (PostgreSQL)
 * Stewart Smith (MySQL)
 * Dan Kegal (Wine)
 * Ben Collins (Ubuntu / Kernel)
 * Jason Kridner (Texas Instruments)
 * Jeremy Allison (Samba / Google)
 * Christian Hammond (VMWare)
 * Ian McKeller (Songbird)
 * Alison Randall (Parrot / Perl / OSCON)
 * David Huffman (LVM)
 * Brian Will (Pigeon)
 * Belinda Lopez (Ubuntu)
 * Ilan Rabanovich (SoCal Linux Expo)
 * Eddy Mulyono (Packaging)
 * Matthew Walster (Demo Scene)

The show will also involve a large exhibition area with a range of
projects, organisations and companies with plenty to see and do. The
first batch of confirmed exhibitors (at the show there will be another
twice as many exhibitors) include:

 * Google
 * Dice
 * GNOME
 * PostgreSQL
 * O'Reilly
 * OpenSuSE
 * Linden Labs
 * Magnatune
 * gOS
 * Neuros
 * Sun
 * Texas Instruments
 * South California Linux Expo
 * Komputers 4 R Kids
 * San Francisco LUG
 * BytesFree.org
 * Ontario Linux Fest
 * Frets on Fire
 * OpenNMS
 * One Course Source
 * Haiku
 * Hyperic
 * BitRock
 * Ubuntu
 * OpenStreetMap
 * The Digital Tipping Point
 * FreeBSD
 * Inkscape

In addition to this, the LugRadio team will be recording a live
performance of their cult-hit podcast, which has over 20,000
listeners, in front of the LugRadio Live USA 2008 audience - like the
UK event, this is always quite a spectacle, and excellent fun for all
involved.

More information about LugRadio Live USA 2008 can be found at
http://lugradio.org/live/USA2008/

--------

Thanks!

  Jono and the LugRadio Live USA 2008 team

-------------------------------------------------------

_______________________________________________
Oakland mailing list
Oakland at pm.org
http://mail.pm.org/mailman/listinfo/oakland

----- End forwarded message -----

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From quinn at fairpath.com  Tue Mar 11 12:52:02 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Tue, 11 Mar 2008 11:52:02 -0800
Subject: [sf-perl] Lugradio conference at the Metreon
In-Reply-To: <20080311193224.GA94624@fu.funkspiel.org>
References: <20080311193224.GA94624@fu.funkspiel.org>
Message-ID: <20080311195201.GC94624@fu.funkspiel.org>

On Tue, Mar 11, 2008 at 11:32:24AM -0800, Quinn Weaver wrote:

> Are people up for this?  I don't want to be half-assed about this, but
> if we can get enough people I'm all for it.

And my commitment:  I'll work two slots, but I'd like to keep the other
two open to attend talks.  May add a slot once the talk schedule is
published.

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From fred at redhotpenguin.com  Tue Mar 11 14:08:09 2008
From: fred at redhotpenguin.com (Fred Moyer)
Date: Tue, 11 Mar 2008 14:08:09 -0700
Subject: [sf-perl] Lugradio conference at the Metreon
In-Reply-To: <20080311195201.GC94624@fu.funkspiel.org>
References: <20080311193224.GA94624@fu.funkspiel.org>
	<20080311195201.GC94624@fu.funkspiel.org>
Message-ID: <47D6F4B9.9060001@redhotpenguin.com>

Quinn Weaver wrote:
> On Tue, Mar 11, 2008 at 11:32:24AM -0800, Quinn Weaver wrote:
> 
>> Are people up for this?  I don't want to be half-assed about this, but
>> if we can get enough people I'm all for it.
> 
> And my commitment:  I'll work two slots, but I'd like to keep the other
> two open to attend talks.  May add a slot once the talk schedule is
> published.

Yeah, I'll take a couple slots - there is somewhere to plug a laptop in 
right? :)


From doom at kzsu.stanford.edu  Tue Mar 11 17:01:04 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Tue, 11 Mar 2008 17:01:04 -0700
Subject: [sf-perl] Lugradio conference at the Metreon
In-Reply-To: <20080311193224.GA94624@fu.funkspiel.org>
References: <20080311193224.GA94624@fu.funkspiel.org>
Message-ID: <200803120001.m2C014YF005605@kzsu.stanford.edu>


Quinn Weaver <quinn at fairpath.com> wrote:

> Wow.  LugRadio is putting on a Linux/open-source conference April 12 and 13
> at the Metreon.  Registration is $10 - Google and Dice are picking up
> most of the tab.  They have an impressive list of speakers.
>
> It's funny, just yesterday I was mailing Fred about LinuxWorld
> preparation (I want to get started now; it's in August).
>
> One month is a pretty tight schedule, but we could pull off a booth
> for LugRadio iff we get enough boothies.  The conference is on a
> weekend--is anyone up to helping?  I need as many people as possible
> to run four-hour shifts; we need two people at the booth at any
> given time.  So that's probably eight slots to fill, total.
> Call it ten, in case someone needs to cancel at the last minute.

I can take at least two.

(Has anyone heard anything from the codecon folks?  Their web site is
seems a little too inactive for an event that's really happening.
I tried shooting them some email, but got no response.)


From quinn at fairpath.com  Tue Mar 11 18:23:46 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Tue, 11 Mar 2008 17:23:46 -0800
Subject: [sf-perl] Lugradio conference at the Metreon
In-Reply-To: <200803120001.m2C014YF005605@kzsu.stanford.edu>
References: <20080311193224.GA94624@fu.funkspiel.org>
	<200803120001.m2C014YF005605@kzsu.stanford.edu>
Message-ID: <20080312012346.GB96056@fu.funkspiel.org>

On Tue, Mar 11, 2008 at 05:01:04PM -0700, Joe Brenner wrote:

> (Has anyone heard anything from the codecon folks?  Their web site is
> seems a little too inactive for an event that's really happening.
> I tried shooting them some email, but got no response.)

I'm on their announce list, but I haven't seen anything.

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From josh at agliodbs.com  Tue Mar 11 19:29:01 2008
From: josh at agliodbs.com (Josh Berkus)
Date: Tue, 11 Mar 2008 19:29:01 -0700
Subject: [sf-perl] Lugradio conference at the Metreon
In-Reply-To: <200803120001.m2C014YF005605@kzsu.stanford.edu>
References: <20080311193224.GA94624@fu.funkspiel.org>
	<200803120001.m2C014YF005605@kzsu.stanford.edu>
Message-ID: <47D73FED.5090500@agliodbs.com>

Joe, Quinn,

I think PostgresQL will have a booth at LugRadio.

--Josh

From quinn at fairpath.com  Wed Mar 12 09:20:40 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Wed, 12 Mar 2008 08:20:40 -0800
Subject: [sf-perl] Lugradio conference at the Metreon
In-Reply-To: <47D73FED.5090500@agliodbs.com>
References: <20080311193224.GA94624@fu.funkspiel.org>
	<200803120001.m2C014YF005605@kzsu.stanford.edu>
	<47D73FED.5090500@agliodbs.com>
Message-ID: <20080312162040.GA5051@fu.funkspiel.org>

On Tue, Mar 11, 2008 at 07:29:01PM -0700, Josh Berkus wrote:
> Joe, Quinn,
> 
> I think PostgresQL will have a booth at LugRadio.

Excellent.

For those keeping score at home:  I have just under a week to
apply for the booth.  We have ~6 slots taken care of, so that means we
need just 4 more filled to meet our minimum.  (Reminder:  a slot is a
four-hour commitment to hang out at the booth, with at least one other
person.  This is either Saturday or Sunday, April 12 or 13, during the day.)

So, if we can get four more people to work one shift apiece, or two more people
to work two shifts apiece, we're home-free.  And we can get on to signage,
literature, etc., which we'll have a month to handle.

Volunteers?

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From rdm at cfcl.com  Wed Mar 12 22:08:43 2008
From: rdm at cfcl.com (Rich Morin)
Date: Wed, 12 Mar 2008 21:08:43 -0800
Subject: [sf-perl] Fwd: [BALUG-Talk] March 15 OLPC meet at SF State
Message-ID: <p0623090ec3fe6724851a@[192.168.1.205]>

This could be fun...

-r


 --- Begin Forward ---

 Date: Wed, 12 Mar 2008 12:11:13 -0700
 From: Sameer Verma <sverma at sfsu.edu>
 To: balug-talk-balug.org at lists.balug.org
 Subject: [BALUG-Talk] March 15 OLPC meet at SF State

 Note: Also on the web at http://opensource.sfsu.edu/node/479

 Calling all San Francisco Bay Area OLPC enthusiasts!

 If you have an XO laptop, or are just interested in seeing how they
 work, come on down to SF State campus for a get together on March 15,
 2008. If you have an XO or two (or ten), bring them! It is after all St.
 Patrick's Day and the XO is very green indeed!

 Don't have an XO? Download a LiveCD from
 http://dev.laptop.org/pub/livebackupcd/ or simply bring your enthusiasm.

 What: OLPC meet at San Francisco State University

 Why: Build a mesh network, share activities, generate crazy ideas. You
 can sit around the campus (we are not that large) and mesh up with
 others. How far will the mesh go? Only one way to find out! Bring a
 jacket if you plan on sitting outdoors. This is, after all, San
 Francisco weather!

 When: Saturday, March 15, 2008 from 10am to 2pm

 Where: Business Bldg. 202, SF State main campus (1600 Holloway Ave. San
 Francisco, CA 94132).

     Map: http://www.sfsu.edu/~sfsumap/

     By car: http://www.sfsu.edu/~parking/directions/main_campus/car.html

     By Muni: http://www.sfsu.edu/~parking/directions/main_campus/muni.html

 Who: Sponsored by IMSA Student Association (http://imsa.sfsu.edu/).


 Sameer

 --
 Dr. Sameer Verma, Ph.D.
 Associate Professor of Information Systems
 San Francisco State University
 San Francisco CA 94132 USA
 http://verma.sfsu.edu/
 http://opensource.sfsu.edu/

 _______________________________________________
 BALUG-Talk mailing list
 BALUG-Talk at lists.balug.org
 http://lists.balug.org/listinfo.cgi/balug-talk-balug.org

 --- End Forward ---

-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From duane.obrien at gmail.com  Wed Mar 12 21:23:54 2008
From: duane.obrien at gmail.com (Duane Obrien)
Date: Wed, 12 Mar 2008 21:23:54 -0700
Subject: [sf-perl] Fwd: [BALUG-Talk] March 15 OLPC meet at SF State
In-Reply-To: <p0623090ec3fe6724851a@192.168.1.205>
References: <p0623090ec3fe6724851a@192.168.1.205>
Message-ID: <d7127f0b0803122123g5b3b435bt24c357f4c3fa81a0@mail.gmail.com>

I'd love to go.  If mine ever got here.  I'm one of those guys who
fell through the cracks and won't get theirs until next week at the
earliest.

has anyone else on list gotten one?

On Wed, Mar 12, 2008 at 10:08 PM, Rich Morin <rdm at cfcl.com> wrote:
> This could be fun...
>
>  -r
>
>
>   --- Begin Forward ---
>
>   Date: Wed, 12 Mar 2008 12:11:13 -0700
>   From: Sameer Verma <sverma at sfsu.edu>
>   To: balug-talk-balug.org at lists.balug.org
>   Subject: [BALUG-Talk] March 15 OLPC meet at SF State
>
>   Note: Also on the web at http://opensource.sfsu.edu/node/479
>
>   Calling all San Francisco Bay Area OLPC enthusiasts!
>
>   If you have an XO laptop, or are just interested in seeing how they
>   work, come on down to SF State campus for a get together on March 15,
>   2008. If you have an XO or two (or ten), bring them! It is after all St.
>   Patrick's Day and the XO is very green indeed!
>
>   Don't have an XO? Download a LiveCD from
>   http://dev.laptop.org/pub/livebackupcd/ or simply bring your enthusiasm.
>
>   What: OLPC meet at San Francisco State University
>
>   Why: Build a mesh network, share activities, generate crazy ideas. You
>   can sit around the campus (we are not that large) and mesh up with
>   others. How far will the mesh go? Only one way to find out! Bring a
>   jacket if you plan on sitting outdoors. This is, after all, San
>   Francisco weather!
>
>   When: Saturday, March 15, 2008 from 10am to 2pm
>
>   Where: Business Bldg. 202, SF State main campus (1600 Holloway Ave. San
>   Francisco, CA 94132).
>
>      Map: http://www.sfsu.edu/~sfsumap/
>
>      By car: http://www.sfsu.edu/~parking/directions/main_campus/car.html
>
>      By Muni: http://www.sfsu.edu/~parking/directions/main_campus/muni.html
>
>   Who: Sponsored by IMSA Student Association (http://imsa.sfsu.edu/).
>
>
>   Sameer
>
>   --
>   Dr. Sameer Verma, Ph.D.
>   Associate Professor of Information Systems
>   San Francisco State University
>   San Francisco CA 94132 USA
>   http://verma.sfsu.edu/
>   http://opensource.sfsu.edu/
>
>   _______________________________________________
>   BALUG-Talk mailing list
>   BALUG-Talk at lists.balug.org
>   http://lists.balug.org/listinfo.cgi/balug-talk-balug.org
>
>   --- End Forward ---
>
>  --
>  http://www.cfcl.com/rdm            Rich Morin
>  http://www.cfcl.com/rdm/resume     rdm at cfcl.com
>  http://www.cfcl.com/rdm/weblog     +1 650-873-7841
>
>  Technical editing and writing, programming, and web development
>  _______________________________________________
>  SanFrancisco-pm mailing list
>  SanFrancisco-pm at pm.org
>  http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>



-- 
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Don't ever tell anybody anything. If you do, you start missing everybody.

From rdm at cfcl.com  Wed Mar 12 22:38:23 2008
From: rdm at cfcl.com (Rich Morin)
Date: Wed, 12 Mar 2008 21:38:23 -0800
Subject: [sf-perl] Fwd: [BALUG-Talk] March 15 OLPC meet at SF State
In-Reply-To: <d7127f0b0803122123g5b3b435bt24c357f4c3fa81a0@mail.gmail.com>
References: <p0623090ec3fe6724851a@192.168.1.205>
	<d7127f0b0803122123g5b3b435bt24c357f4c3fa81a0@mail.gmail.com>
Message-ID: <p0623090fc3fe6df71ec1@[192.168.1.205]>

At 21:23 -0700 3/12/08, Duane Obrien wrote:
> has anyone else on list gotten one?

V& I and got ours a while back.  We tend to bring it to BASS
and other XO owners have been showing up with them.

  http://cfcl.com/rdm/bass/

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From matisse at matisse.net  Thu Mar 13 10:52:55 2008
From: matisse at matisse.net (Matisse Enzer)
Date: Thu, 13 Mar 2008 10:52:55 -0700
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
Message-ID: <F2B20D23-9C36-4548-8167-9E8F24CDC53A@matisse.net>


On Mar 5, 2008, at 3:46 PM, nheller at silcon.com wrote:
>
>    if ($action =~ /start/i) {     <-------- line 511, the offender
>        &MainPage($q, \%p);
>    }
>    elsif ($action =~ /new/i ||
>           $action =~ /del/i ||
>           $action =~ /edt/i ||
>           $action =~ /save/i )
>    {
>
>


I wonder if you should also change this to:



    my $main_sub = sub { MainPage( $q,\%p };
    my $other_sub  = sub { do_stuff( @_ ) };

    my %dispatch_to = (
        start     => $main_sub,
        new       => $other_sub,
        del       => $other_sub,
        edit      => $other_sub,
        save      => $other_sub,
    );

    my $lower_case_action = lc $action;
    my $dispatch_sub = $dispatch_to{ $lower_case_action };

    $dispatch_sub->( @other_args_can_go_here );


-------------------------------------------------------
Matisse Enzer <matisse at matisse.net>
http://www.matisse.net/  - http://www.eigenstate.net/




From biztos at mac.com  Thu Mar 13 11:27:27 2008
From: biztos at mac.com (frosty)
Date: Thu, 13 Mar 2008 11:27:27 -0700
Subject: [sf-perl] Can't figure this one out
In-Reply-To: <F2B20D23-9C36-4548-8167-9E8F24CDC53A@matisse.net>
References: <1602.66.93.171.131.1204760811.squirrel@webmail.silcon.com>
	<F2B20D23-9C36-4548-8167-9E8F24CDC53A@matisse.net>
Message-ID: <1D3EC9A8-0118-1000-8774-5664DAB756B9-Webmail-10021@mac.com>

...or even:

    my $main_sub = sub { MainPage( $q,\%p ) };


;-)

-- frosty
 
On Thursday, March 13, 2008, at 10:53AM, "Matisse Enzer" <matisse at matisse.net> wrote:
>I wonder if you should also change this to:
>
>    my $main_sub = sub { MainPage( $q,\%p };
>    ...etc...

From rdm at cfcl.com  Thu Mar 13 19:20:02 2008
From: rdm at cfcl.com (Rich Morin)
Date: Thu, 13 Mar 2008 18:20:02 -0800
Subject: [sf-perl] Fwd: [BALUG-Announce] Mark Shuttleworth at BALUG
 2008-03-25!, and other BALUG News
Message-ID: <p06230914c3ff90c046b3@[192.168.1.205]>

I realize this falls on top of the normal SFPUG evening,
but I don't get a chance to hear Mark Shuttleworth very
often.  Hmmmm; anyone want to have an SFPUG table at the
dinner?

-r


 --- Begin Forward ---

 Date: Thu, 13 Mar 2008 07:34:26 -0700
 From: Michael Paoli <Michael.Paoli at cal.berkeley.edu>
 To: balug-announce at lists.balug.org
 Subject: [BALUG-Announce] Mark Shuttleworth at BALUG 2008-03-25!,
 	and other BALUG News

 Mark Shuttleworth at BALUG 2008-03-25!, and other BALUG News

 Contents:
 * Mark Shuttleworth at BALUG 2008-03-25 (note change from usual date)
 * Additional upcoming BALUG meeting dates and speakers
 * LugRadio Live USA 2008 in San Francisco
 * Contacts, etc.

 ------------------------------

 Mark Shuttleworth at BALUG 2008-03-25 (note change from usual date)

 NOTE SPECIAL MEETING DATE: 2008-03-25 (no meeting on 2008-03-18)
 For our 2008-03-25 meeting we are proud to present:

 Mark Shuttleworth[1] of Ubuntu[2]/Canonical[3].

 Mark is founder of the Ubuntu Project[2], and also founded
 Thawte[4] (which specialized in digital certificates),
 HBD Venture Capital[5] (invests in early stage South African companies
 with high growth potential) and
 The Shuttleworth Foundation[6] (non-profit organization that accelerates
 social innovation in Africa with a particular focus on education), and
 also formed Canonical[3] (commercial sponsor of Ubuntu, and committed to
 the development, distribution and promotion of open source software
 products, and to providing tools and support to the open source
 community).  In April 2002, Mark also became the
 First African in Space[7], arriving via Soyuz TM-34 to the
 International Space Station and spending several days there, and
 returning via Soyuz TM-33.

 If you intend to come, please RSVP to: rsvp at balug.org
 Why RSVP?  Well, we probably wouldn't turn you away, but the RSVPs
 really help the Four Seas Restaurant plan settings/food/staffing and
 help ensure that we're able to eat upstairs in the private banquet room.

 Subscribe to our (low volume) Announce list[8] to be sure and be
 notified of our upcoming events.

 1. http://www.markshuttleworth.com/biography
 2. http://www.ubuntu.com/
 3. http://www.canonical.com/
 4. http://www.thawte.com/
 5. http://www.hbd.com/
 6. http://www.shuttleworthfoundation.org/
 7. http://www.africaninspace.com/
 8. http://www.new.balug.org/#Lists

 ------------------------------

 Additional upcoming BALUG meeting dates[9] and speakers (see our web
 page[9] for more details):
 2008-04-15 Eric Allman
 2008-05-20 Jeremy Allison
 2008-06-17 Andrew Morton
 2008-07-15 Mike Linksvayer, CTO, Creative Commons
 Subscribe to our (quite low volume) Announce list[8] to be sure and be
 notified of our upcoming events.

 8. http://www.new.balug.org/#Lists
 9. http://www.new.balug.org/#Meetings%20(upcoming)

 ------------------------------

 LugRadio Live USA 2008[10] at The Metreon in San Francisco, April 12th &
 13th, 2008

 LugRadio Live USA 2008 brings San Francisco the unique atmosphere of
 LugRadio Live UK, an event that has developed a strong reputation for
 providing a range of topics about free software, Open Source, digital
 rights, technology and more, a compelling list of speakers, exhibitors
 and birds of a feather sessions, and wrapping it all in a unique, fun,
 loose, social and inclusive event, which is often described as
 combining the atmosphere of a rock concert and a computer conference.

 BALUG may also be participating in this - along with other User Groups
 in the area.

 10. http://www.lugradio.org/live/USA2008/

 ------------------------------

 Contacting BALUG: http://www.new.balug.org/#Contact
 Feedback on our publicity/announcements (e.g. contacts or lists where we
 should get our information out that we're not presently reaching, or
 things we should do differently): publicity-feedback at balug.org
 _______________________________________________
 BALUG-Announce mailing list
 BALUG-Announce at lists.balug.org
 http://lists.balug.org/listinfo.cgi/balug-announce-balug.org

 --- End Forward ---

-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From quinn at fairpath.com  Thu Mar 13 18:35:30 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Thu, 13 Mar 2008 17:35:30 -0800
Subject: [sf-perl] Fwd: [BALUG-Announce] Mark Shuttleworth at BALUG
	2008-03-25!, and other BALUG News
In-Reply-To: <p06230914c3ff90c046b3@[192.168.1.205]>
References: <p06230914c3ff90c046b3@[192.168.1.205]>
Message-ID: <20080314013530.GB20761@fu.funkspiel.org>

On Thu, Mar 13, 2008 at 06:20:02PM -0800, Rich Morin wrote:
> I realize this falls on top of the normal SFPUG evening,
> but I don't get a chance to hear Mark Shuttleworth very
> often.  Hmmmm; anyone want to have an SFPUG table at the
> dinner?

Absolutely.  And this is a good stand-in, IMO, for a March meeting.

I have a meeting in the works for April 22, by the way.  Details to follow
when stuff in confirmed.

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From afife at untangle.com  Fri Mar 14 12:26:26 2008
From: afife at untangle.com (Andrew Fife)
Date: Fri, 14 Mar 2008 12:26:26 -0700 (PDT)
Subject: [sf-perl] Fwd: [BALUG-Announce] Mark Shuttleworth at BALUG
In-Reply-To: <mailman.27.1205521226.21757.sanfrancisco-pm@pm.org>
References: <mailman.27.1205521226.21757.sanfrancisco-pm@pm.org>
Message-ID: <006d01c88609$669249f0$0200a8c0@Untangle.local>

>Hmmmm; anyone want to have an SFPUG table at the
>dinner?

I'd be happy to setup a SFPM table, just remember to bring a placard so
that everyone knows where it is.  Incidentally, the Sonoma PM group is
also making the Shuttleworth event their March Meeting.

----------------------------------------
Andrew Fife
Untangle - Open Source Security Gateway
download.untangle.com

650.425.3327 (O)
415.806.6028 (C)
afife at untangle.com 



From extasia at extasia.org  Fri Mar 14 19:38:47 2008
From: extasia at extasia.org (David Alban)
Date: Fri, 14 Mar 2008 19:38:47 -0700
Subject: [sf-perl] parallelize system tasks, collecting statuses and output
Message-ID: <4c714a9c0803141938o327847d7gc7419856c45a1514@mail.gmail.com>

greetings,

our sysadmins regularly need to restart a service on, say thirty
machines.  the service takes two and a half minutes to restart.  i'd
like to write a perl program they can use to parallelize the restart.
so that the whole operation takes three minutes rather than an hour or
more.  i want to collect the statuses and any output of the service
restart commands.

i found the Bidirectional Communication with Yourself section of the
perlipc man page.  i'm trying to hack their example so that only the
child writer writes to the parent reader.  my hacked version forks two
child processes, which i want to run in parallel.  each child process
ssh's to a host, sleeps a small amount of time, and then runs the
hostname command.  but the child procs seem to be running serially.

#!/usr/bin/perl

use warnings;
use strict;

     # log_timestamp() below comes from this module
use <LOCAL LOGGING MODULE>;

use IO::Handle;

my @hosts = qw( hostname_1 hostname_2 );
my $numprocs = @hosts;

my @output;

for my $instance ( 1..$numprocs ) {
  my $index = $instance - 1;

  pipe( CHILD_READER, PARENT_WRITER )
    or die "can't pipe( CHILD_READER, PARENT_WRITER ): $!\n";

  PARENT_WRITER->autoflush( 1 );

  my $pid;

       # parent
  if ( $pid = fork() ) {
    close PARENT_WRITER;

    chomp( my $line = <CHILD_READER> );
    $output[ $index ] = $line;

    close CHILD_READER;
  } # if

       # child
  else {
    not defined $pid and die "can't fork: $!\n";

    close CHILD_READER;

    my $host = $hosts[ $index ];
    my @results = qx{ ssh $host "sleep 5; hostname" };

    print PARENT_WRITER
          log_timestamp(),
          " child pid $$; instance $instance; results => ",
          join( '', @results );

    close PARENT_WRITER;

    exit;
  } # if
} # for

for my $instance ( 1..$numprocs ) {
  my $index = $instance - 1;
  print $output[ $index ], "\n";
} # if



--------------------------------------------------------------------------------

i execute this as:

  $ date; perl junk; date


and get:

  Sat Mar 15 02:30:22 UTC 2008
  2008-03-15 02:30:27 +0000 child pid 5469; instance 1; results => hostname_1
  2008-03-15 02:30:32 +0000 child pid 5471; instance 2; results => hostname_2
  Sat Mar 15 02:30:32 UTC 2008


it's running the second child process only after the first one
finishes, which defeats my goal of parallelizing.  what am i missing?

there's so much stuff out there that promises to help with this and i
don't know whether i'm going down the wrong path.  surely you fine
folks must have done stuff like this before.  do indeed tell me to
rtfm, but please tell me which fm (or other doc) to r.

please also feel free to tell me i'm going about it totally the wrong
way, perhaps with a pointer in the general (better) direction.

thanks,
david
-- 
Live in a world of your own, but always welcome visitors.

From friedman at highwire.stanford.edu  Fri Mar 14 22:12:09 2008
From: friedman at highwire.stanford.edu (Michael Friedman)
Date: Fri, 14 Mar 2008 22:12:09 -0700
Subject: [sf-perl] parallelize system tasks,
	collecting statuses and output
In-Reply-To: <4c714a9c0803141938o327847d7gc7419856c45a1514@mail.gmail.com>
References: <4c714a9c0803141938o327847d7gc7419856c45a1514@mail.gmail.com>
Message-ID: <0D87E766-1BEC-4A0A-A6FD-2C2F43E7B674@highwire.stanford.edu>

David,

I think the sequentialism is in your code. You fork() in a loop, and  
each child does its thing, but the parent process -- the one that is  
about to fork the next child -- is sitting there waiting on the  
child's output:

	chomp( my $line = <CHILD_READER> );

Until the child gives it something, you've got blocking IO there. It  
won't start the next child until after it's received the output from  
the previous child. To make it actually parallel, you have to not wait  
on output from the child until after you've created all the children.  
You can do that via a double-fork server, some sort of signaling, or...

Have you considered using threads? Perl now comes with real live  
working thread support (if you compiled it in). The threads API is  
pretty brilliant, IMO, as it makes it almost as easy as in Java to  
write multi-threaded code.

http://search.cpan.org/~jdhedden/threads-1.69/threads.pm

If you didn't happen to compile in thread support, or are running on  
someone else's copy of perl who didn't, there's a CPAN module that  
emulates the thread API using fork(). It's much slower than real  
threads, but lets you write code that works without threads now and  
with them later when you can update perl.

http://search.cpan.org/~rybskej/forks-0.27/lib/forks.pm

In either case, it explicitly lets you detach from threads and let  
them run at their own pace, gathering the results later. Your driving  
script can only pick up the results in single file, but threads may  
finish in a different order than you started them. I've used this a  
couple of times and it's pretty cool. ;-)

I think the threads API would be a good solution to your problem, if  
it's available to you.

-- Mike


On Mar 14, 2008, at 7:38 PM, David Alban wrote:

> greetings,
>
> our sysadmins regularly need to restart a service on, say thirty
> machines.  the service takes two and a half minutes to restart.  i'd
> like to write a perl program they can use to parallelize the restart.
> so that the whole operation takes three minutes rather than an hour or
> more.  i want to collect the statuses and any output of the service
> restart commands.
>
> i found the Bidirectional Communication with Yourself section of the
> perlipc man page.  i'm trying to hack their example so that only the
> child writer writes to the parent reader.  my hacked version forks two
> child processes, which i want to run in parallel.  each child process
> ssh's to a host, sleeps a small amount of time, and then runs the
> hostname command.  but the child procs seem to be running serially.
>
> #!/usr/bin/perl
>
> use warnings;
> use strict;
>
>     # log_timestamp() below comes from this module
> use <LOCAL LOGGING MODULE>;
>
> use IO::Handle;
>
> my @hosts = qw( hostname_1 hostname_2 );
> my $numprocs = @hosts;
>
> my @output;
>
> for my $instance ( 1..$numprocs ) {
>  my $index = $instance - 1;
>
>  pipe( CHILD_READER, PARENT_WRITER )
>    or die "can't pipe( CHILD_READER, PARENT_WRITER ): $!\n";
>
>  PARENT_WRITER->autoflush( 1 );
>
>  my $pid;
>
>       # parent
>  if ( $pid = fork() ) {
>    close PARENT_WRITER;
>
>    chomp( my $line = <CHILD_READER> );
>    $output[ $index ] = $line;
>
>    close CHILD_READER;
>  } # if
>
>       # child
>  else {
>    not defined $pid and die "can't fork: $!\n";
>
>    close CHILD_READER;
>
>    my $host = $hosts[ $index ];
>    my @results = qx{ ssh $host "sleep 5; hostname" };
>
>    print PARENT_WRITER
>          log_timestamp(),
>          " child pid $$; instance $instance; results => ",
>          join( '', @results );
>
>    close PARENT_WRITER;
>
>    exit;
>  } # if
> } # for
>
> for my $instance ( 1..$numprocs ) {
>  my $index = $instance - 1;
>  print $output[ $index ], "\n";
> } # if
>
>
>
> --------------------------------------------------------------------------------
>
> i execute this as:
>
>  $ date; perl junk; date
>
>
> and get:
>
>  Sat Mar 15 02:30:22 UTC 2008
>  2008-03-15 02:30:27 +0000 child pid 5469; instance 1; results =>  
> hostname_1
>  2008-03-15 02:30:32 +0000 child pid 5471; instance 2; results =>  
> hostname_2
>  Sat Mar 15 02:30:32 UTC 2008
>
>
> it's running the second child process only after the first one
> finishes, which defeats my goal of parallelizing.  what am i missing?
>
> there's so much stuff out there that promises to help with this and i
> don't know whether i'm going down the wrong path.  surely you fine
> folks must have done stuff like this before.  do indeed tell me to
> rtfm, but please tell me which fm (or other doc) to r.
>
> please also feel free to tell me i'm going about it totally the wrong
> way, perhaps with a pointer in the general (better) direction.
>
> thanks,
> david
> -- 
> Live in a world of your own, but always welcome visitors.
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

---------------------------------------------------------------------
Michael Friedman                     HighWire Press
Phone: 650-725-1974                  Stanford University
FAX:   270-721-8034                  <friedman at highwire.stanford.edu>
---------------------------------------------------------------------



From biztos at mac.com  Fri Mar 14 23:39:23 2008
From: biztos at mac.com (Kevin Frost)
Date: Fri, 14 Mar 2008 23:39:23 -0700
Subject: [sf-perl] parallelize system tasks,
	collecting statuses and output
In-Reply-To: <0D87E766-1BEC-4A0A-A6FD-2C2F43E7B674@highwire.stanford.edu>
References: <4c714a9c0803141938o327847d7gc7419856c45a1514@mail.gmail.com>
	<0D87E766-1BEC-4A0A-A6FD-2C2F43E7B674@highwire.stanford.edu>
Message-ID: <458154C4-C25B-449E-B06B-2E445EB58E59@mac.com>

This sounds like a fun project, but I'm pretty sure there are industry- 
standard tools that already solve this problem.

I don't know what they use, but our guys do this all the time on  
hundreds of boxes.  Just FYI.  I'd still do it in perl. :-)

  -- frosty

(via iPhone)


On Mar 14, 2008, at 10:12 PM, Michael Friedman <friedman at highwire.stanford.edu 
 > wrote:

> David,
>
> I think the sequentialism is in your code. You fork() in a loop, and
> each child does its thing, but the parent process -- the one that is
> about to fork the next child -- is sitting there waiting on the
> child's output:
>
>    chomp( my $line = <CHILD_READER> );
>
> Until the child gives it something, you've got blocking IO there. It
> won't start the next child until after it's received the output from
> the previous child. To make it actually parallel, you have to not wait
> on output from the child until after you've created all the children.
> You can do that via a double-fork server, some sort of signaling,  
> or...
>
> Have you considered using threads? Perl now comes with real live
> working thread support (if you compiled it in). The threads API is
> pretty brilliant, IMO, as it makes it almost as easy as in Java to
> write multi-threaded code.
>
> http://search.cpan.org/~jdhedden/threads-1.69/threads.pm
>
> If you didn't happen to compile in thread support, or are running on
> someone else's copy of perl who didn't, there's a CPAN module that
> emulates the thread API using fork(). It's much slower than real
> threads, but lets you write code that works without threads now and
> with them later when you can update perl.
>
> http://search.cpan.org/~rybskej/forks-0.27/lib/forks.pm
>
> In either case, it explicitly lets you detach from threads and let
> them run at their own pace, gathering the results later. Your driving
> script can only pick up the results in single file, but threads may
> finish in a different order than you started them. I've used this a
> couple of times and it's pretty cool. ;-)
>
> I think the threads API would be a good solution to your problem, if
> it's available to you.
>
> -- Mike
>
>
> On Mar 14, 2008, at 7:38 PM, David Alban wrote:
>
>> greetings,
>>
>> our sysadmins regularly need to restart a service on, say thirty
>> machines.  the service takes two and a half minutes to restart.  i'd
>> like to write a perl program they can use to parallelize the restart.
>> so that the whole operation takes three minutes rather than an hour  
>> or
>> more.  i want to collect the statuses and any output of the service
>> restart commands.
>>
>> i found the Bidirectional Communication with Yourself section of the
>> perlipc man page.  i'm trying to hack their example so that only the
>> child writer writes to the parent reader.  my hacked version forks  
>> two
>> child processes, which i want to run in parallel.  each child process
>> ssh's to a host, sleeps a small amount of time, and then runs the
>> hostname command.  but the child procs seem to be running serially.
>>
>> #!/usr/bin/perl
>>
>> use warnings;
>> use strict;
>>
>>    # log_timestamp() below comes from this module
>> use <LOCAL LOGGING MODULE>;
>>
>> use IO::Handle;
>>
>> my @hosts = qw( hostname_1 hostname_2 );
>> my $numprocs = @hosts;
>>
>> my @output;
>>
>> for my $instance ( 1..$numprocs ) {
>> my $index = $instance - 1;
>>
>> pipe( CHILD_READER, PARENT_WRITER )
>>   or die "can't pipe( CHILD_READER, PARENT_WRITER ): $!\n";
>>
>> PARENT_WRITER->autoflush( 1 );
>>
>> my $pid;
>>
>>      # parent
>> if ( $pid = fork() ) {
>>   close PARENT_WRITER;
>>
>>   chomp( my $line = <CHILD_READER> );
>>   $output[ $index ] = $line;
>>
>>   close CHILD_READER;
>> } # if
>>
>>      # child
>> else {
>>   not defined $pid and die "can't fork: $!\n";
>>
>>   close CHILD_READER;
>>
>>   my $host = $hosts[ $index ];
>>   my @results = qx{ ssh $host "sleep 5; hostname" };
>>
>>   print PARENT_WRITER
>>         log_timestamp(),
>>         " child pid $$; instance $instance; results => ",
>>         join( '', @results );
>>
>>   close PARENT_WRITER;
>>
>>   exit;
>> } # if
>> } # for
>>
>> for my $instance ( 1..$numprocs ) {
>> my $index = $instance - 1;
>> print $output[ $index ], "\n";
>> } # if
>>
>>
>>
>> --- 
>> --- 
>> --- 
>> --- 
>> --------------------------------------------------------------------
>>
>> i execute this as:
>>
>> $ date; perl junk; date
>>
>>
>> and get:
>>
>> Sat Mar 15 02:30:22 UTC 2008
>> 2008-03-15 02:30:27 +0000 child pid 5469; instance 1; results =>
>> hostname_1
>> 2008-03-15 02:30:32 +0000 child pid 5471; instance 2; results =>
>> hostname_2
>> Sat Mar 15 02:30:32 UTC 2008
>>
>>
>> it's running the second child process only after the first one
>> finishes, which defeats my goal of parallelizing.  what am i missing?
>>
>> there's so much stuff out there that promises to help with this and i
>> don't know whether i'm going down the wrong path.  surely you fine
>> folks must have done stuff like this before.  do indeed tell me to
>> rtfm, but please tell me which fm (or other doc) to r.
>>
>> please also feel free to tell me i'm going about it totally the wrong
>> way, perhaps with a pointer in the general (better) direction.
>>
>> thanks,
>> david
>> -- 
>> Live in a world of your own, but always welcome visitors.
>> _______________________________________________
>> SanFrancisco-pm mailing list
>> SanFrancisco-pm at pm.org
>> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
> ---------------------------------------------------------------------
> Michael Friedman                     HighWire Press
> Phone: 650-725-1974                  Stanford University
> FAX:   270-721-8034                  <friedman at highwire.stanford.edu>
> ---------------------------------------------------------------------
>
>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm

From merlyn at stonehenge.com  Sat Mar 15 08:16:08 2008
From: merlyn at stonehenge.com (Randal L. Schwartz)
Date: Sat, 15 Mar 2008 08:16:08 -0700
Subject: [sf-perl] parallelize system tasks,
	collecting statuses and output
In-Reply-To: <4c714a9c0803141938o327847d7gc7419856c45a1514@mail.gmail.com>
	(David Alban's message of "Fri, 14 Mar 2008 19:38:47 -0700")
References: <4c714a9c0803141938o327847d7gc7419856c45a1514@mail.gmail.com>
Message-ID: <86d4pwyq4n.fsf@blue.stonehenge.com>

>>>>> "David" == David Alban <extasia at extasia.org> writes:

David> our sysadmins regularly need to restart a service on, say thirty
David> machines.  the service takes two and a half minutes to restart.  i'd
David> like to write a perl program they can use to parallelize the restart.
David> so that the whole operation takes three minutes rather than an hour or
David> more.  i want to collect the statuses and any output of the service
David> restart commands.

David> i found the Bidirectional Communication with Yourself section of the
David> perlipc man page.  i'm trying to hack their example so that only the
David> child writer writes to the parent reader.  my hacked version forks two
David> child processes, which i want to run in parallel.  each child process
David> ssh's to a host, sleeps a small amount of time, and then runs the
David> hostname command.  but the child procs seem to be running serially.

"Getting your kids to do the work"
   http://www.stonehenge.com/merlyn/LinuxMag/col15.html

Or, just look at Parallel::ForkManager in the CPAN.

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn at stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Perl/Unix/security consulting, Technical writing, Comedy, etc. etc.
See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!

From lavendula6654 at yahoo.com  Mon Mar 17 18:50:46 2008
From: lavendula6654 at yahoo.com (Elaine)
Date: Mon, 17 Mar 2008 18:50:46 -0700 (PDT)
Subject: [sf-perl] Software Development Courses at Foothill College
Message-ID: <14913.45835.qm@web31708.mail.mud.yahoo.com>

Spring quarter classes start Monday, 7 April, at
Foothill College.
These two may be of interest to you:

1) Application Software Development with Ajax - 5
units
Prerequisite: Knowledge  of HTML and JavaScript 
COIN 71 - Tuesday evenings at Middlefield campus in
Palo Alto

2) Introduction to Python Programming - 5 units
Prerequisite: Any programming language experience
CIS 68K - Monday evenings at Middlefield campus in
Palo Alto

If you are interested in taking a class, please
register as soon as possible by going to:
http://www.foothill.fhda.edu/reg/index.php
If not enough students sign up, a class may be
cancelled.

If you have any questions, please contact the
instructor, Elaine Haight, at
haightElaine at foothill.edu




      ____________________________________________________________________________________
Looking for last minute shopping deals?  
Find them fast with Yahoo! Search.  http://tools.search.yahoo.com/newsearch/category.php?category=shopping

From x.sfpug at chatterjee.net  Mon Mar 17 23:28:54 2008
From: x.sfpug at chatterjee.net (Anirvan Chatterjee)
Date: Mon, 17 Mar 2008 23:28:54 -0700
Subject: [sf-perl] Advice on how to expose a class-wide preference
Message-ID: <20080318062854.GA4496@chatterjee.net>

I'm looking for advice on how to expose a user-settable preference
that affects the behavior of all objects in a given class. I'm working
on a module representing data stringifiable in multiple ways, e.g.:

	$data->normalized
	$data->full
	$data->compact

I'd like to set up string operator overloading, but it's not obvious
what the default stringification style should be; different users will
have different needs. I'd like to allow users to choose how objects
will stringify by default in their code. I can imagine doing this in
two different ways:

1) Have a class-wide preference indicating the default style:

	Class->default_string_style('normalized');

2) Let users set their preference via a magic variable:

	local $Class::default_string_style = 'normalized';

The first option seems cleaner, and easier to subclass (were that an
issue); it also allows for better error handling when the user calls
the class method. Unfortunately, setting a class-wide flag also blows
away any prior settings. Two pieces of code using the module will
necessarily overwrite each other's preferences.

Using a variable seems uglier, but allows users to safely intermix
different preference values in different contexts.

Any suggestions on which of these is generally better to use? Are
there other strategies I'm ignoring? Thanks for any advice.

-- 
Anirvan Chatterjee
www.chatterjee.net
www.bookfinder.com

From shlomif at iglu.org.il  Tue Mar 18 01:50:12 2008
From: shlomif at iglu.org.il (Shlomi Fish)
Date: Tue, 18 Mar 2008 10:50:12 +0200
Subject: [sf-perl] Advice on how to expose a class-wide preference
In-Reply-To: <20080318062854.GA4496@chatterjee.net>
References: <20080318062854.GA4496@chatterjee.net>
Message-ID: <200803181050.12869.shlomif@iglu.org.il>

Hi Anirvan!

On Tuesday 18 March 2008, Anirvan Chatterjee wrote:
> I'm looking for advice on how to expose a user-settable preference
> that affects the behavior of all objects in a given class. I'm working
> on a module representing data stringifiable in multiple ways, e.g.:
>
> 	$data->normalized
> 	$data->full
> 	$data->compact
>
> I'd like to set up string operator overloading, but it's not obvious
> what the default stringification style should be; different users will
> have different needs. I'd like to allow users to choose how objects
> will stringify by default in their code. I can imagine doing this in
> two different ways:
>
> 1) Have a class-wide preference indicating the default style:
>
> 	Class->default_string_style('normalized');
>
> 2) Let users set their preference via a magic variable:
>
> 	local $Class::default_string_style = 'normalized';
>
> The first option seems cleaner, and easier to subclass (were that an
> issue); it also allows for better error handling when the user calls
> the class method. Unfortunately, setting a class-wide flag also blows
> away any prior settings. Two pieces of code using the module will
> necessarily overwrite each other's preferences.
>
> Using a variable seems uglier, but allows users to safely intermix
> different preference values in different contexts.
>
> Any suggestions on which of these is generally better to use? Are
> there other strategies I'm ignoring? Thanks for any advice.

Another option I can think of is to make use of closures to define a certain 
scope:

<<<<<<<<<<<<<<<<<<<

$data->as_normalized(sub {
	# Here $data will stringify as normalized:
	print "$date";
});
# And here $data will revert back to its previous setting.

>>>>>>>>>>>>>>>>>>>

And naturally inside the closure you can safely set the setting to any other 
value, because it will be reverted at the end.

Regards,

	Shlomi Fish

---------------------------------------------------------------------
Shlomi Fish      shlomif at iglu.org.il
Homepage:        http://www.shlomifish.org/

I'm not an actor - I just play one on T.V.

From moseley at hank.org  Tue Mar 18 04:44:43 2008
From: moseley at hank.org (Bill Moseley)
Date: Tue, 18 Mar 2008 04:44:43 -0700
Subject: [sf-perl] Advice on how to expose a class-wide preference
In-Reply-To: <20080318062854.GA4496@chatterjee.net>
References: <20080318062854.GA4496@chatterjee.net>
Message-ID: <20080318114442.GD6464@hank.org>

On Mon, Mar 17, 2008 at 11:28:54PM -0700, Anirvan Chatterjee wrote:
> I'm looking for advice on how to expose a user-settable preference
> that affects the behavior of all objects in a given class. I'm working
> on a module representing data stringifiable in multiple ways, e.g.:

Are the objects created in many different places?  That is, would it
be that difficult to just specify the type of stringification in the
object?


    my $data = Thing->new(
        stringification_method => $config->stringification_method,
    );

If you go the class data route I'd still copy the preference into the
object if you might ever want to change the behavior of some objects.

-- 
Bill Moseley
moseley at hank.org


From woof at danlo.com  Tue Mar 18 06:23:23 2008
From: woof at danlo.com (Daniel Lo)
Date: Tue, 18 Mar 2008 06:23:23 -0700
Subject: [sf-perl] Advice on how to expose a class-wide preference
In-Reply-To: <20080318062854.GA4496@chatterjee.net>
References: <20080318062854.GA4496@chatterjee.net>
Message-ID: <199058546.20080318062323@danlo.com>

Hello Anirvan,

        It appears to me that you are talking about a User Interface
description.  I don't know of any reason for a class API to return different
results depending on a normalized, full, or compact method. (At least I can't
think of one right now. :)  Your code should never very in its return values
other wise you will be writing different APIs for each object/function to handle
receiving $data->normalized() $data->full() $data->compact() and $data->human()
$data->and_so_on.  Things get exponentially worse after that.

        If you are debating on how to return content for the end user, you
should always return $data->full().  Then the templating code or the display
code should be responsible for taking $data->full() and making it normalized,
compact or full.

        You should consider a visitor pattern or using Template::Toolkit (very
good btw).  http://en.wikipedia.org/wiki/Visitor_pattern.

        This is great for deciding how you want to display the data.

        You may want to check out Class::Prototyped ( very cool :)

http://search.cpan.org/~teverett/Class-Prototyped-1.11/lib/Class/Prototyped.pm

-daniel




Monday, March 17, 2008, 11:28:54 PM, you wrote:

> I'm looking for advice on how to expose a user-settable preference
> that affects the behavior of all objects in a given class. I'm working
> on a module representing data stringifiable in multiple ways, e.g.:

>         $data->normalized
>         $data->full
>         $data->compact

> I'd like to set up string operator overloading, but it's not obvious
> what the default stringification style should be; different users will
> have different needs. I'd like to allow users to choose how objects
> will stringify by default in their code. I can imagine doing this in
> two different ways:

> 1) Have a class-wide preference indicating the default style:

>         Class->default_string_style('normalized');

> 2) Let users set their preference via a magic variable:

>         local $Class::default_string_style = 'normalized';

> The first option seems cleaner, and easier to subclass (were that an
> issue); it also allows for better error handling when the user calls
> the class method. Unfortunately, setting a class-wide flag also blows
> away any prior settings. Two pieces of code using the module will
> necessarily overwrite each other's preferences.

> Using a variable seems uglier, but allows users to safely intermix
> different preference values in different contexts.

> Any suggestions on which of these is generally better to use? Are
> there other strategies I'm ignoring? Thanks for any advice.




-- 
Best regards,
 Daniel                            mailto:woof at danlo.com


From sphink at gmail.com  Tue Mar 18 13:12:46 2008
From: sphink at gmail.com (Steve Fink)
Date: Tue, 18 Mar 2008 13:12:46 -0700
Subject: [sf-perl] Advice on how to expose a class-wide preference
In-Reply-To: <20080318062854.GA4496@chatterjee.net>
References: <20080318062854.GA4496@chatterjee.net>
Message-ID: <7d7f2e8c0803181312l79bc3d80lea00f6daecf3e683@mail.gmail.com>

On Mon, Mar 17, 2008 at 11:28 PM, Anirvan Chatterjee
<x.sfpug at chatterjee.net> wrote:
>  1) Have a class-wide preference indicating the default style:
>
>         Class->default_string_style('normalized');
>
>  2) Let users set their preference via a magic variable:
>
>         local $Class::default_string_style = 'normalized';
>
>  The first option seems cleaner, and easier to subclass (were that an
>  issue); it also allows for better error handling when the user calls
>  the class method. Unfortunately, setting a class-wide flag also blows
>  away any prior settings. Two pieces of code using the module will
>  necessarily overwrite each other's preferences.
>
>  Using a variable seems uglier, but allows users to safely intermix
>  different preference values in different contexts.

Just to add to the set of possibilities, you could make the settings
only apply to the package that requested them. So the syntax could
either be your original #1:

  Class->default_string_style('normalized')

or you could make it explicit

  Class->default_string_style(__PACKAGE__, 'normalized')

default_string_style() would then use caller() or the passed-in
package name to record who wants what, and the stringification would
use caller() to decide whose setting to use.

I'm not recommending it, just pointing it out.

If you really wanted to be able to use local() but still have
package-scoping, you could always do

  our $STRINGIFICATION_SETTING;
  local $STRINGIFICATION_SETTING = ...;

and have stringification use caller() to look up that variable in its
caller's package.

Whee.

From fluxnet at gmail.com  Tue Mar 18 14:37:07 2008
From: fluxnet at gmail.com (Bern)
Date: Tue, 18 Mar 2008 14:37:07 -0700
Subject: [sf-perl] Advice on how to expose a class-wide preference
In-Reply-To: <7d7f2e8c0803181312l79bc3d80lea00f6daecf3e683@mail.gmail.com>
References: <20080318062854.GA4496@chatterjee.net>
	<7d7f2e8c0803181312l79bc3d80lea00f6daecf3e683@mail.gmail.com>
Message-ID: <f4b0d8050803181437k2de9a473ib25fde9aa01d4a7@mail.gmail.com>

hi, sorry need to ask when is the next meeting ? or any other meetings and
or events in the sfbay area, thanks

On Tue, Mar 18, 2008 at 1:12 PM, Steve Fink <sphink at gmail.com> wrote:

> On Mon, Mar 17, 2008 at 11:28 PM, Anirvan Chatterjee
> <x.sfpug at chatterjee.net> wrote:
> >  1) Have a class-wide preference indicating the default style:
> >
> >         Class->default_string_style('normalized');
> >
> >  2) Let users set their preference via a magic variable:
> >
> >         local $Class::default_string_style = 'normalized';
> >
> >  The first option seems cleaner, and easier to subclass (were that an
> >  issue); it also allows for better error handling when the user calls
> >  the class method. Unfortunately, setting a class-wide flag also blows
> >  away any prior settings. Two pieces of code using the module will
> >  necessarily overwrite each other's preferences.
> >
> >  Using a variable seems uglier, but allows users to safely intermix
> >  different preference values in different contexts.
>
> Just to add to the set of possibilities, you could make the settings
> only apply to the package that requested them. So the syntax could
> either be your original #1:
>
>  Class->default_string_style('normalized')
>
> or you could make it explicit
>
>  Class->default_string_style(__PACKAGE__, 'normalized')
>
> default_string_style() would then use caller() or the passed-in
> package name to record who wants what, and the stringification would
> use caller() to decide whose setting to use.
>
> I'm not recommending it, just pointing it out.
>
> If you really wanted to be able to use local() but still have
> package-scoping, you could always do
>
>  our $STRINGIFICATION_SETTING;
>  local $STRINGIFICATION_SETTING = ...;
>
> and have stringification use caller() to look up that variable in its
> caller's package.
>
> Whee.
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.pm.org/pipermail/sanfrancisco-pm/attachments/20080318/48ce9eff/attachment.html 

From quinn at fairpath.com  Tue Mar 18 16:38:51 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Tue, 18 Mar 2008 15:38:51 -0800
Subject: [sf-perl] Advice on how to expose a class-wide preference
In-Reply-To: <f4b0d8050803181437k2de9a473ib25fde9aa01d4a7@mail.gmail.com>
References: <20080318062854.GA4496@chatterjee.net>
	<7d7f2e8c0803181312l79bc3d80lea00f6daecf3e683@mail.gmail.com>
	<f4b0d8050803181437k2de9a473ib25fde9aa01d4a7@mail.gmail.com>
Message-ID: <20080318233851.GA75876@fu.funkspiel.org>

On Tue, Mar 18, 2008 at 02:37:07PM -0700, Bern wrote:
> hi, sorry need to ask when is the next meeting ? or any other meetings and
> or events in the sfbay area, thanks

My bad--I haven't marked this up to the Web site yet.

Our next meeting is March 25.  We will join BALUG in a dinner with
Mark Shuttleworth, the founder of Canonical--the company behind Ubuntu.
(Mark is also famous as the founder of Thawte and as a space tourist.)

Andrew Fife of BALUG has kindly agreed to set aside a table for SF.pm.
If you will be attending, please mail me.  I need to get Andrew
an accurate RSVP count.

For more details, see http://www.balug.org/ (scroll down to "Next meetings:").
I will post this on sf.pm.org soon.

Thanks!

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From fluxnet at gmail.com  Tue Mar 18 20:31:14 2008
From: fluxnet at gmail.com (Bern)
Date: Tue, 18 Mar 2008 20:31:14 -0700
Subject: [sf-perl] Advice on how to expose a class-wide preference
In-Reply-To: <20080318233851.GA75876@fu.funkspiel.org>
References: <20080318062854.GA4496@chatterjee.net>
	<7d7f2e8c0803181312l79bc3d80lea00f6daecf3e683@mail.gmail.com>
	<f4b0d8050803181437k2de9a473ib25fde9aa01d4a7@mail.gmail.com>
	<20080318233851.GA75876@fu.funkspiel.org>
Message-ID: <f4b0d8050803182031l28e76cf1k41c2c2b44e8bbb5f@mail.gmail.com>

mail u here  to sanfrancisco-pm at pm.org or ?
please consider me a sure thing

On Tue, Mar 18, 2008 at 4:38 PM, Quinn Weaver <quinn at fairpath.com> wrote:

> On Tue, Mar 18, 2008 at 02:37:07PM -0700, Bern wrote:
> > hi, sorry need to ask when is the next meeting ? or any other meetings
> and
> > or events in the sfbay area, thanks
>
> My bad--I haven't marked this up to the Web site yet.
>
> Our next meeting is March 25.  We will join BALUG in a dinner with
> Mark Shuttleworth, the founder of Canonical--the company behind Ubuntu.
> (Mark is also famous as the founder of Thawte and as a space tourist.)
>
> Andrew Fife of BALUG has kindly agreed to set aside a table for SF.pm.
> If you will be attending, please mail me.  I need to get Andrew
> an accurate RSVP count.
>
> For more details, see http://www.balug.org/ (scroll down to "Next
> meetings:").
> I will post this on sf.pm.org soon.
>
> Thanks!
>
> --
> Quinn Weaver, independent contractor  |  President, San Francisco Perl
> Mongers
> http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
> 510-520-5217 <http://sf.pm.org/510-520-5217>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.pm.org/pipermail/sanfrancisco-pm/attachments/20080318/481375dc/attachment.html 

From pm.org at daveola.com  Wed Mar 19 17:07:45 2008
From: pm.org at daveola.com (David Ljung Madison)
Date: Wed, 19 Mar 2008 17:07:45 -0700
Subject: [sf-perl] Porting to Windows and writing a GUI?  (ugh)
Message-ID: <E1Jc8Xr-0007vg-Kv@dave.getdave.com>

I have a shareware script that I've mostly ported to Windows
via CygWin (and ActivePerl, at one point).

I wanted to write/get a GUI front-end to call the script
(and maybe an installer as well) so I can make it a little
more user-friendly for the non-perl users out there.

I don't really do Windows myself, so I might just hire
someone to make the front-end (whether in perl or not),
but it occurred to me that someone may have already solved
this problem in the general sense.

It would be snifty, it seems to me, if there was a GUI
that could call a script and could be customized for
some of the command-line flags, or, even better, could
perhaps read them from the script itself.

Any suggestions?

And if not - any thoughts on where I could hire someone
for a small job like this?

Dave
Root
---------------------------------------------------------------------------
Dave Ljung Madison          http://GetDave.com/                415.341.5555
--- I am totally into Perl now.  I love it.  It rules with an iron fist. --

From doom at kzsu.stanford.edu  Wed Mar 19 18:54:18 2008
From: doom at kzsu.stanford.edu (Joe Brenner)
Date: Wed, 19 Mar 2008 18:54:18 -0700
Subject: [sf-perl] Porting to Windows and writing a GUI? (ugh)
In-Reply-To: <E1Jc8Xr-0007vg-Kv@dave.getdave.com>
References: <E1Jc8Xr-0007vg-Kv@dave.getdave.com>
Message-ID: <200803200154.m2K1sIfc060398@kzsu.stanford.edu>


David Ljung Madison <pm.org at daveola.com> wrote:

> I have a shareware script that I've mostly ported to Windows
> via CygWin (and ActivePerl, at one point).
>
> I wanted to write/get a GUI front-end to call the script
> (and maybe an installer as well) so I can make it a little
> more user-friendly for the non-perl users out there.

There are a few existing toolkits that could be used to do something
like this.  Tk is the traditional one, but lately people have been
talking about Wx, which I believe is supposed to look a little better
(or more like a native app):

  http://search.cpan.org/dist/Wx/

If you try this, don't forget to check the Install file:

  "You need to install wxWidgets before you can compile wxPerl."



From ddascalescu at gmail.com  Thu Mar 20 20:12:31 2008
From: ddascalescu at gmail.com (Dan Dascalescu)
Date: Thu, 20 Mar 2008 20:12:31 -0700
Subject: [sf-perl] Porting to Windows and writing a GUI? (ugh)
In-Reply-To: <200803200154.m2K1sIfc060398@kzsu.stanford.edu>
References: <E1Jc8Xr-0007vg-Kv@dave.getdave.com>
	<200803200154.m2K1sIfc060398@kzsu.stanford.edu>
Message-ID: <3561cc6d0803202012x635ba8c8r25a32ab332b24a5@mail.gmail.com>

I've found Win32::GUI much easier to comprehend, use and get started with.

Dan

On Wed, Mar 19, 2008 at 6:54 PM, Joe Brenner <doom at kzsu.stanford.edu> wrote:
>
>  David Ljung Madison <pm.org at daveola.com> wrote:
>
>  > I have a shareware script that I've mostly ported to Windows
>  > via CygWin (and ActivePerl, at one point).
>  >
>  > I wanted to write/get a GUI front-end to call the script
>  > (and maybe an installer as well) so I can make it a little
>  > more user-friendly for the non-perl users out there.
>
>  There are a few existing toolkits that could be used to do something
>  like this.  Tk is the traditional one, but lately people have been
>  talking about Wx, which I believe is supposed to look a little better
>  (or more like a native app):
>
>   http://search.cpan.org/dist/Wx/
>
>  If you try this, don't forget to check the Install file:
>
>   "You need to install wxWidgets before you can compile wxPerl."

From quinn at fairpath.com  Mon Mar 24 09:21:16 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Mon, 24 Mar 2008 08:21:16 -0800
Subject: [sf-perl] Reminder:  Mark Shuttleworth tomorrow.  RSVP!
Message-ID: <20080324162116.GA10796@fu.funkspiel.org>

Reminder: Mark Shuttleworth is speaking at BALUG tomorrow night.
BALUG has offered to give us a special table, but we need an RSVP
count.  So far I've received only one.

If you want to join us, please mail me at quinn at fairpath.com (don't reply
to the list) and I'll get you on the guest list.  If you don't want to
sit with us, you don't have to RSVP, but it does help the BALUG folks.

Thanks a lot,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From rdm at cfcl.com  Tue Mar 25 08:53:51 2008
From: rdm at cfcl.com (Rich Morin)
Date: Tue, 25 Mar 2008 07:53:51 -0800
Subject: [sf-perl] Freeway Threeway - three evening of goodness!
Message-ID: <p0623094ac40ecffe24d7@[192.168.1.205]>

Even if you've been hiding under the covers, waiting for things
to quiet down a bit, you're probably aware of the week's events:

  Tue: BALUG (San Francisco) - Mark Shuttleworth on Ubuntu, etc.

    -- http://www.new.balug.org

  Wed: BASS (San Francisco)  - the usual Beer, Pizza, and chatter

    -- http://www.cfcl.com/rdm/bass

  Thu: PC&P (Redwood City)   - pizza and a double feature on RSpec

    --
<http://ruby.meetup.com/123/calendar/7190591/>http://ruby.meetup.com/123/calendar/7190591/

Be there or be elsewhere...

-r
-- 
http://www.cfcl.com/rdm            Rich Morin
http://www.cfcl.com/rdm/resume     rdm at cfcl.com
http://www.cfcl.com/rdm/weblog     +1 650-873-7841

Technical editing and writing, programming, and web development

From quinn at fairpath.com  Tue Mar 25 18:09:16 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Tue, 25 Mar 2008 17:09:16 -0800
Subject: [sf-perl] Reminder:  Mark Shuttleworth tomorrow.  RSVP!
In-Reply-To: <20080324162116.GA10796@fu.funkspiel.org>
References: <20080324162116.GA10796@fu.funkspiel.org>
Message-ID: <20080326010916.GB38456@fu.funkspiel.org>

BALUG update:

Thanks for the RSVPs.  We have seven people confirmed.  There should
be a table with a sign just for us.

Unfortunately, and to my immense regret, I won't be there.  I realized
at the last minute that I double-booked tonight.  Double duh.  The other
event can't happen without me, so I have to drop BALUG.

Needless to say, I am really bummed.  I would be enormously grateful
if someone could ask Mark this question, in some form (write it down
or paraphrase, I don't care which):

    What would it take for Kubuntu or Ubuntu to become as popular as,
    or more popular than, Windows and Macs on the desktop?

I know the answer of many people on this list would be "It'll never
happen" or "The Mac OS X GUI is way better."  But I want Mark's
opinion. :)

Thanks, and see you next time,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From quinn at fairpath.com  Fri Mar 28 12:10:35 2008
From: quinn at fairpath.com (Quinn Weaver)
Date: Fri, 28 Mar 2008 11:10:35 -0800
Subject: [sf-perl] LugRadio is on
Message-ID: <20080328191035.GA71716@fu.funkspiel.org>

And now for some pleasant news.

We've reached our staffing goals, and LugRadio has confirmed us for a booth.
We're on!

So far we have Joe, Fred, Yary, and me covering shifts.  I'm going to start an
off-list thread to handle logistics.  If you want to, it's not too late
to pitch in; email me (now or later) and I'll add you to that thread.

For the record, here are the tasks that need doing:

- Next two weeks:  print a sign.  I'll take point on this.

- Next two weeks:  help us set up an awesome Perl demo app.
  Brian's Google mashup would be good, if he wants to bring it again.
  Ideally, I'd also like to have something that doesn't depend on the
  wifi network (conferences' shared wifi is notoriously flaky).

- Friday evening, April 10:  load-in and booth setup (sign, books, computer[s]).

- Saturday:  booth staffing

- Sunday:  booth staffing

- Sunday afternoon:  teardown

That's it!  Hope to see you at the conference, and/or on the other thread. :)

Thanks,

-- 
Quinn Weaver, independent contractor  |  President, San Francisco Perl Mongers
http://fairpath.com/quinn/resume/     |  http://sf.pm.org/
510-520-5217

From nheller at silcon.com  Fri Mar 28 12:44:28 2008
From: nheller at silcon.com (Neil Heller)
Date: Fri, 28 Mar 2008 12:44:28 -0700
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <20080328191035.GA71716@fu.funkspiel.org>
References: <20080328191035.GA71716@fu.funkspiel.org>
Message-ID: <005001c8910c$22f2c630$68d85290$@com>


An interesting question came up.

Suppose you have a large number (> 250k) of strings in an array.
You wish to eliminate duplicates.

Is it faster to sort the array and then iterate once through the list
(writing selected ones to a new array) or write them all into a hash (at
which time duplicates would be naturally eliminated)?

Neil Heller



From fluxnet at gmail.com  Fri Mar 28 12:49:43 2008
From: fluxnet at gmail.com (Bern)
Date: Fri, 28 Mar 2008 12:49:43 -0700
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <005001c8910c$22f2c630$68d85290$@com>
References: <20080328191035.GA71716@fu.funkspiel.org>
	<005001c8910c$22f2c630$68d85290$@com>
Message-ID: <f4b0d8050803281249s34a20c02l332a3662191a3eb3@mail.gmail.com>

hash will never be faster then nothing (at least me thinks)
you can benchmark them tho

On Fri, Mar 28, 2008 at 12:44 PM, Neil Heller <nheller at silcon.com> wrote:

>
> An interesting question came up.
>
> Suppose you have a large number (> 250k) of strings in an array.
> You wish to eliminate duplicates.
>
> Is it faster to sort the array and then iterate once through the list
> (writing selected ones to a new array) or write them all into a hash (at
> which time duplicates would be naturally eliminated)?
>
> Neil Heller
>
>
> _______________________________________________
> SanFrancisco-pm mailing list
> SanFrancisco-pm at pm.org
> http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.pm.org/pipermail/sanfrancisco-pm/attachments/20080328/26460eee/attachment.html 

From andy at petdance.com  Fri Mar 28 12:52:16 2008
From: andy at petdance.com (Andy Lester)
Date: Fri, 28 Mar 2008 14:52:16 -0500
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <005001c8910c$22f2c630$68d85290$@com>
References: <20080328191035.GA71716@fu.funkspiel.org>
	<005001c8910c$22f2c630$68d85290$@com>
Message-ID: <4B3D7CEC-A9C9-4025-AF31-5309929AB06D@petdance.com>


On Mar 28, 2008, at 2:44 PM, Neil Heller wrote:
> Is it faster to sort the array and then iterate once through the list
> (writing selected ones to a new array) or write them all into a hash  
> (at
> which time duplicates would be naturally eliminated)?


It's easy enough to test it and not have to speculate idly.

--
Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance





From nheller at silcon.com  Fri Mar 28 13:15:44 2008
From: nheller at silcon.com (Neil Heller)
Date: Fri, 28 Mar 2008 13:15:44 -0700
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <4B3D7CEC-A9C9-4025-AF31-5309929AB06D@petdance.com>
References: <20080328191035.GA71716@fu.funkspiel.org>	<005001c8910c$22f2c630$68d85290$@com>
	<4B3D7CEC-A9C9-4025-AF31-5309929AB06D@petdance.com>
Message-ID: <006b01c89110$8110df50$83329df0$@com>

It's easy to fill an array with 250,0000 or more strings?

Neil Heller
510-862-4387


-----Original Message-----
From: sanfrancisco-pm-bounces+nheller=silcon.com at pm.org
[mailto:sanfrancisco-pm-bounces+nheller=silcon.com at pm.org] On Behalf Of Andy
Lester
Sent: Friday, March 28, 2008 12:52 PM
To: San Francisco Perl Mongers User Group
Subject: Re: [sf-perl] Sorting vs. hash


On Mar 28, 2008, at 2:44 PM, Neil Heller wrote:
> Is it faster to sort the array and then iterate once through the list
> (writing selected ones to a new array) or write them all into a hash  
> (at
> which time duplicates would be naturally eliminated)?


It's easy enough to test it and not have to speculate idly.

--
Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance




_______________________________________________
SanFrancisco-pm mailing list
SanFrancisco-pm at pm.org
http://mail.pm.org/mailman/listinfo/sanfrancisco-pm


From andy at petdance.com  Fri Mar 28 13:18:51 2008
From: andy at petdance.com (Andy Lester)
Date: Fri, 28 Mar 2008 15:18:51 -0500
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <006b01c89110$8110df50$83329df0$@com>
References: <20080328191035.GA71716@fu.funkspiel.org>	<005001c8910c$22f2c630$68d85290$@com>
	<4B3D7CEC-A9C9-4025-AF31-5309929AB06D@petdance.com>
	<006b01c89110$8110df50$83329df0$@com>
Message-ID: <224D33B1-CD23-469E-A62A-81293A3CD84D@petdance.com>


On Mar 28, 2008, at 3:15 PM, Neil Heller wrote:
> It's easy to fill an array with 250,0000 or more strings?


Sure.

--
Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance





From biztos at mac.com  Fri Mar 28 13:22:45 2008
From: biztos at mac.com (frosty)
Date: Fri, 28 Mar 2008 13:22:45 -0700
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <006b01c89110$8110df50$83329df0$@com>
References: <20080328191035.GA71716@fu.funkspiel.org>
	<005001c8910c$22f2c630$68d85290$@com>
	<4B3D7CEC-A9C9-4025-AF31-5309929AB06D@petdance.com>
	<006b01c89110$8110df50$83329df0$@com>
Message-ID: <519061F2-0118-1000-9552-CDA1EBCAFC30-Webmail-10019@mac.com>

Sure.  See below.  I think the sorting trick is fastest.

use strict;
use warnings;
use Benchmark;

# starting point: one big fat array (Neil says 250K).
# but let's assume there are some dupes.
my @input;
my @ltr        = ( 'a' .. 'z' );
my $dupe_count = 0;
while ( @input < 250000 ) {
    my $t = join( '', map { $ltr[ rand(@ltr) ] } ( 1 .. rand(40) + 10 ) );
    push( @input, $t );

    # Eliminating dupes is the point.
    if ( @input / 20 == int( @input / 20 ) ) {
        unshift( @input, $t );
        $dupe_count++;
    }
}
printf "%d strings (%d dupes)\n", scalar(@input), $dupe_count;

timethese(
    100,
    {

        SortStyle  => \&array_strip_without_hash,
        SeenStyle  => \&array_strip_with_hash,
        SimpleHash => \&hash_strip,
        SortedHash => \&hash_strip_sort
    }
);

sub array_strip_without_hash {

    # This strikes me as really dumb, but...
    my @new_array;
    my $last_item = '';
    foreach my $item ( sort @input ) {
        next if $item eq $last_item;
        $last_item = $item;
        push( @new_array, $item );

    }
    die "oops" unless @new_array == @input - $dupe_count;

}

sub array_strip_with_hash {

    # My thinking is this is also dumb.
    my @new_array;
    my %seen;
    foreach my $item (@input) {
        next if $seen{$item};
        push( @new_array, $item );
        $seen{$item} = 1;
    }
    die "oops" unless @new_array == @input - $dupe_count;

}

sub hash_strip {

    my %new_hash = map { $_ => 1 } @input;
    my @new_array = keys %new_hash;
    die "oops" unless @new_array == @input - $dupe_count;

}

sub hash_strip_sort {

    # because it's more useful.
    my %new_hash = map { $_ => 1 } @input;
    my @new_array = sort keys %new_hash;
    die "oops" unless @new_array == @input - $dupe_count;
}

 
On Friday, March 28, 2008, at 01:16PM, "Neil Heller" <nheller at silcon.com> wrote:
>It's easy to fill an array with 250,0000 or more strings?
>
>Neil Heller
>510-862-4387
>
>
>-----Original Message-----
>From: sanfrancisco-pm-bounces+nheller=silcon.com at pm.org
>[mailto:sanfrancisco-pm-bounces+nheller=silcon.com at pm.org] On Behalf Of Andy
>Lester
>Sent: Friday, March 28, 2008 12:52 PM
>To: San Francisco Perl Mongers User Group
>Subject: Re: [sf-perl] Sorting vs. hash
>
>
>On Mar 28, 2008, at 2:44 PM, Neil Heller wrote:
>> Is it faster to sort the array and then iterate once through the list
>> (writing selected ones to a new array) or write them all into a hash  
>> (at
>> which time duplicates would be naturally eliminated)?
>
>
>It's easy enough to test it and not have to speculate idly.
>
>--
>Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance
>
>
>
>
>_______________________________________________
>SanFrancisco-pm mailing list
>SanFrancisco-pm at pm.org
>http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
>_______________________________________________
>SanFrancisco-pm mailing list
>SanFrancisco-pm at pm.org
>http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
>

From gregc at cnet.com  Fri Mar 28 13:26:39 2008
From: gregc at cnet.com (greg lee coleman [1482])
Date: Fri, 28 Mar 2008 13:26:39 -0700 (PDT)
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <519061F2-0118-1000-9552-CDA1EBCAFC30-Webmail-10019@mac.com>
References: <20080328191035.GA71716@fu.funkspiel.org>
	<005001c8910c$22f2c630$68d85290$@com>
	<4B3D7CEC-A9C9-4025-AF31-5309929AB06D@petdance.com>
	<006b01c89110$8110df50$83329df0$@com>
	<519061F2-0118-1000-9552-CDA1EBCAFC30-Webmail-10019@mac.com>
Message-ID: <Pine.LNX.4.64.0803281325230.881@c17-nasa-techweb.cnet.com>


i know he's lurking, but anyway - 

http://en.wikipedia.org/wiki/Schwartzian_transform


hth,
-gleeco

From biztos at mac.com  Fri Mar 28 13:28:38 2008
From: biztos at mac.com (frosty)
Date: Fri, 28 Mar 2008 13:28:38 -0700
Subject: [sf-perl] Sorting vs. hash
In-Reply-To: <519061F2-0118-1000-9552-CDA1EBCAFC30-Webmail-10019@mac.com>
References: <20080328191035.GA71716@fu.funkspiel.org>
	<005001c8910c$22f2c630$68d85290$@com>
	<4B3D7CEC-A9C9-4025-AF31-5309929AB06D@petdance.com>
	<006b01c89110$8110df50$83329df0$@com>
	<519061F2-0118-1000-9552-CDA1EBCAFC30-Webmail-10019@mac.com>
Message-ID: <519061F2-0118-1000-9572-CDA1EBCAFC30-Webmail-10019@mac.com>

D'oh, no!  I take it back!  That'll teach me to not wait for my benchmarks to run.

It looks like the good old %seen trick is fastest on my MacBook Pro (sysperl, 5.8.8).

The hash via map stuff is a lot slower.  If I had more time to play with it I'd like to know which parts make it slower but I have to get back to work.

My hunch is that if you really cared, you could make the sort thingy the fastest by using a different sort algo.

-- f.

On Friday, March 28, 2008, at 01:23PM, "frosty" <biztos at mac.com> wrote:
>Sure.  See below.  I think the sorting trick is fastest.
>
>use strict;
>use warnings;
>use Benchmark;
>
># starting point: one big fat array (Neil says 250K).
># but let's assume there are some dupes.
>my @input;
>my @ltr        = ( 'a' .. 'z' );
>my $dupe_count = 0;
>while ( @input < 250000 ) {
>    my $t = join( '', map { $ltr[ rand(@ltr) ] } ( 1 .. rand(40) + 10 ) );
>    push( @input, $t );
>
>    # Eliminating dupes is the point.
>    if ( @input / 20 == int( @input / 20 ) ) {
>        unshift( @input, $t );
>        $dupe_count++;
>    }
>}
>printf "%d strings (%d dupes)\n", scalar(@input), $dupe_count;
>
>timethese(
>    100,
>    {
>
>        SortStyle  => \&array_strip_without_hash,
>        SeenStyle  => \&array_strip_with_hash,
>        SimpleHash => \&hash_strip,
>        SortedHash => \&hash_strip_sort
>    }
>);
>
>sub array_strip_without_hash {
>
>    # This strikes me as really dumb, but...
>    my @new_array;
>    my $last_item = '';
>    foreach my $item ( sort @input ) {
>        next if $item eq $last_item;
>        $last_item = $item;
>        push( @new_array, $item );
>
>    }
>    die "oops" unless @new_array == @input - $dupe_count;
>
>}
>
>sub array_strip_with_hash {
>
>    # My thinking is this is also dumb.
>    my @new_array;
>    my %seen;
>    foreach my $item (@input) {
>        next if $seen{$item};
>        push( @new_array, $item );
>        $seen{$item} = 1;
>    }
>    die "oops" unless @new_array == @input - $dupe_count;
>
>}
>
>sub hash_strip {
>
>    my %new_hash = map { $_ => 1 } @input;
>    my @new_array = keys %new_hash;
>    die "oops" unless @new_array == @input - $dupe_count;
>
>}
>
>sub hash_strip_sort {
>
>    # because it's more useful.
>    my %new_hash = map { $_ => 1 } @input;
>    my @new_array = sort keys %new_hash;
>    die "oops" unless @new_array == @input - $dupe_count;
>}
>
> 
>On Friday, March 28, 2008, at 01:16PM, "Neil Heller" <nheller at silcon.com> wrote:
>>It's easy to fill an array with 250,0000 or more strings?
>>
>>Neil Heller
>>510-862-4387
>>
>>
>>-----Original Message-----
>>From: sanfrancisco-pm-bounces+nheller=silcon.com at pm.org
>>[mailto:sanfrancisco-pm-bounces+nheller=silcon.com at pm.org] On Behalf Of Andy
>>Lester
>>Sent: Friday, March 28, 2008 12:52 PM
>>To: San Francisco Perl Mongers User Group
>>Subject: Re: [sf-perl] Sorting vs. hash
>>
>>
>>On Mar 28, 2008, at 2:44 PM, Neil Heller wrote:
>>> Is it faster to sort the array and then iterate once through the list
>>> (writing selected ones to a new array) or write them all into a hash  
>>> (at
>>> which time duplicates would be naturally eliminated)?
>>
>>
>>It's easy enough to test it and not have to speculate idly.
>>
>>--
>>Andy Lester => andy at petdance.com => www.petdance.com => AIM:petdance
>>
>>
>>
>>
>>_______________________________________________
>>SanFrancisco-pm mailing list
>>SanFrancisco-pm at pm.org
>>http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>>
>>_______________________________________________
>>SanFrancisco-pm mailing list
>>SanFrancisco-pm at pm.org
>>http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>>
>>
>_______________________________________________
>SanFrancisco-pm mailing list
>SanFrancisco-pm at pm.org
>http://mail.pm.org/mailman/listinfo/sanfrancisco-pm
>
>